Acute and Lifelong Exercise Modulate the Tumorigenic Potential of Human Lung Cancer Cells and Their Susceptibility to Cisplatin

24 The association between higher levels of physical activity and lower cancer risk and mortality is 25 well established. However, a causal link is yet to be proven. Recent studies showed a decrease in the 26 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint proliferation rates of cultured human cancer cells when the human serum employed to stimulate them 27 was conditioned by acute exercise. Here, we tested the hypothesis that serum mediates some of the 28 putative benefits of exercise on cancer through alterations to the growth pattern and susceptibility to 29 chemotherapy agents of cancer cells. To this end, human non -small cell lung cancer (NSCLC) cells 30 were exposed to serum from two cohorts that differed significantly on their levels of physical activity 31 and, accordingly, cardiorespiratory fitness, but were otherwise identical (master athletes and non -32 exercisers), collected before and after an acute exercise intervention. Serum levels of glucose, lipids, 33 albumin, C-reactive protein and cytokines were determined and the impact of the serum responses to 34 acute and lifelong exercise on the above -mentioned parameters were analyzed. We found that acute 35 exercise decreased the cells’ proliferation rate, yet shortened the cells’ lag phase after detachment, 36 whereas lifelong exercise had the opposite effects. Significantly, we showed, for the first time, that 37 lifelong exercise increased susceptibility to a chemotherapy agent (cisplatin), which may contribute to 38 the decreased cancer mortality rates found among those who exercise regularly. Similar to the cellular 39 effects, changes to serum cytokine levels – several of them linked to the senescence -associated 40 secretory phenotype – depended on whether serum was conditioned by acute or by chronic exercise. 41

Cancer risk, Cell migration, Cell proliferation, Cytokines, Exercise -conditioned human 42 serum, Susceptibility to chemotherapy 43 44 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

45 Cancer is one of the world’s largest health problems, accounting for every sixth death globally. 46 And yet, a large percentage of cancer cases and cancer deaths could be prevented simply by eliminating 47 or reducing exposure to environmental and occupational risk factors (Urbano et al., 2008; Urbano et 48 al., 2012; Katzke et al., 2015; Lewandowska et al., 2019; Adam et al., 2023) and by adopting healthier 49 lifestyles, namely quitting smoking, maintaining a healthy weight, reducing alcohol consumption, and 50 being more active (Mctiernan, 2008; Friedenreich et al., 2010; Garcia & Pearce, 2023). 51 Up to the early 2000s, robust evidence for an association between physical inactivity and 52 increased cancer risk was restricted to a few site-specific cancers (Vainio et al., 2002; Friedenreich et 53 al., 2010). In the meantime, a wealth of evidence for many other cancer types has been gathered. In 54 particular, a pooled analysis from 2016 of 12 prospective cohort studies involving 1.44 million 55 participants showed that the risk for 13 out of 26 common types of cancer was lower for those 56 participants who self-reported engagement in higher levels of moderate or vigorous intensity leisure -57 time physical activity, than for those participants who self-reported lower levels of engagement (Moore 58 et al. , 2016 ). Specifically, the association was strong (greater than 20% reduction in risk) for 59 oesophageal adenocarcinoma, myeloid leukaemia, and cancers of the liver, lung, kidney, gastric cardia 60 and endometrium, and moderate (10%–20% reduction in risk) for myeloma and cancers of the colon, 61 head and neck, rectum, bladder, and breast(Moore et al., 2016). 62 The evidence for a positive association between exercise and reduced cancer mortality has been 63 very recently strengthened by a randomized trial, conducted at 55 centres and involving a total of 889 64 patients with resected colon cancer, showing that a 3-year structured exercise program initiated soon 65 after adjuvant chemotherapy significantly extended disease-free survival, with findings consistent with 66 longer overall survival (Courneya et al., 2025). 67 Along the same lines, evidence also emerged for a relationship between the amount of exercise 68 and the reduction in the risk of some cancer types (Thune & Furberg, 2001 ). A very recent meta -69 analysis of 96 reported studies involving over 30 million participants yielded an inverse non -linear 70 dose–response association between non -occupational physical activity and disease and mortality 71 outcomes for cancer (total and site -specific), suggesting more pronounced increases in potential 72 benefits from inactive lifestyles up to levels of non -occupational physical activity equivalent to the 73 minimum level recommended by the World Health Organization (WHO) for adults (i.e., “at least 75 74 to 150 minutes of vigorous-intensity aerobic physical activity per week, or at least 150 to 300 minutes 75 of moderate-intensity aerobic physical activity per week, or an equivalent combination throughout the 76 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint week” (Bull et al., 2020)). Benefits became incrementally smaller up to twice the recommended 77 minimum level, the evidence for benefit being weaker thereafter (Garcia & Pearce, 2023). 78 Despite the amount and strength of the evidence just discussed, these associations are not 79 proof that higher levels of physical activity reduce cancer risk and improve cancer management. 80 For instance, it cannot be ruled out that the lower cancer risk found amongst active people is a 81 consequence of other healthy lifestyle behaviours besides being physically active or a result of 82 gene variants enriched in active people. Whether a causal relationship exists can only be 83 ascertained if the cellular and molecular mechanisms behind the putative benefits of active 84 lifestyles to cancer prevention and management are firmly established. 85 It has been argued that the observed reduction of cancer risk and mortality associated with 86 physical activity is primarily an indirect effect of weight loss, especially in terms of adiposity. 87 The fact that an association was found between excess body fat and a higher risk for a large 88 number of cancer types, most notably cancers of digestive organs and cancers of hormone -89 sensitive organs in women (Lauby-Secretan et al., 2016; Avgerinos et al., 2019), lends some 90 support to this view. Lowering the number of adipose cells and/or their size reduces the secretion 91 and, hence, the circulating levels of several biomarkers of cancer risk , namely sex hormones, 92 metabolic hormones ( e.g., leptin and resistin), insulin -like growth factors, pro -inflammatory 93 cytokines, and macrophage infiltration into adipose tissue, ultimately lowering systemic 94 inflammation (Avgerinos et al., 2019; Michailidou et al., 2022; Tilg et al., 2025). In addition, 95 obesity weakens the antitumor responses of natural killer (NK) cells, thus compromising 96 immunosurveillance and the elimination of cancer cells (Michelet et al., 2018). Nonetheless, the 97 fact that the inverse associations between physical activity and cancer risk were unaffected by 98 body size in 10 of the above-mentioned 13 types of common cancers (Moore et al. , 2016 ) 99 suggests that mechanisms independent from weight loss contribute to the putative beneficial 100 effects of physical activity to cancer risk and mortality. 101 There is increasing evidence that the growth pattern s of cultured human cancer cells are 102 altered when the human serum used to stimulate them is conditioned by exercise . These in vitro 103 studies are still sparse and vary considerably in all aspects of study design, from exercise 104 intervention to cohort characteristics, cell lines used, exposure regimen (percentage of serum in 105 the growth medium and duration of the exposure) and assays employed, making direct 106 comparisons difficult. Nonetheless, meta -analyses suggest that , compared to non -conditioned 107 (baseline) serum, serum conditioned by an acute exercise intervention diminishes the viability of 108 cultured cancer cell s, with a large overall effect size that increased with the intensity of the 109 exercise (Orange et al., 2020; Brown et al., 2021; Soares et al., 2021). Of note, no studies on the 110 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint impact of serological responses to exercise on cancer susceptibility to chemotherapy have been 111 reported. In this study, we exposed A549 human non-small cell lung cancer (NSCLC) cells to 112 human sera obtained before or after acute exercise and analyzed the impact of the transient serum 113 responses to the intervention on the sera’s ability to stimulate cancer cell proliferation, on the cells’ 114 plating efficiency (assessed previously in only one study), here used as a metric for reproductive 115 potential, lag phase (not previously reported), cell migration (not previously reported) and 116 susceptibility to chemotherapy (not previously reported). We used two distinct cohorts with different 117 levels of physical activity , volunteers who did not meet the above -mentioned WHO guidelines on 118 physical activity for health (hereafter termed non-exercisers) and master athletes. Through the use of 119 these two cohorts, we were able to also assess the impact of permanent serum changes on the same 120 parameters, as well as to gain insight into how the impact of acute exercise on serum’s cancer cell 121 modulatory properties depends on levels of physical activity throughout life. 122 123 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

124 Study overview and ethics statement 125 Twenty-six Caucasian healthy male volunteers, aged 40 to 63 years were recruited based on 126 their levels of physical activity and assigned to two age -matched cohorts : 13 non-exercisers 127 (individuals who self-reported not having met in the 20 years that preceded their recruitment the 128 WHO guidelines on physical activity for health (Bull et al., 2020)) and 13 age-matched master 129 athletes, all of them self-reported experienced competitors who had been training and competing 130 for at least 20 years in their sports modality at the time of recruitment. Blood was collected from 131 each participant before (baseline serum) and after the exercise intervention (post-exercise serum). 132 Serum prepared from the collected blood was used to evaluate the effects of serum responses to 133 acute and lifelong exercise on the pattern of growth of human lung cancer cells, providing 134 valuable information regarding their transformation degree and, ultimately, their tumorigenic 135 potential, as well as on the susceptibility of these cells to the chemotherap y agent cisplatin. All 136 sera were tested in parallel. For ethical considerations, the amount of blood collected was limited, 137 meaning that pooled sera, instead of individual sera, were used in the assessment of plating 138 efficiencies, susceptibly to cisplatin and cytokine detection. Each of the four pooled sera, 139 corresponding to the four conditions just described, contained equal volumes of sera from 13 140 participants. At the time of recruitment, the intervention and potential risks were explained to the 141 participants who all gave written informed consent before their inclusion in the study. Prior to 142 the acute exercise intervention, participants were subjected to a brief medical examination that 143 included blood pressure measurement , electrocardiogram and a standard medical history 144 questionnaire. Exclusion criteria included smoking, having donated blood in the three months 145 that preceded the intervention, any evidence of chronic or inflammatory diseases, having had an 146 infectious disease up to six weeks prior to the exercise intervention, and taking supplementation 147 or medication. Participants were asked to maintain their lifestyle routines and to avoid caffeine, 148 exercising and partaking in sports activities in the 24 h preceding the exercise intervention. The 149 study was approved by the Ethics Committee of the Faculty of Sports Sciences and Physical 150 Education of the University of Coimbra (reference CE/FCDEF -UC/00062013). All procedures 151 conformed to the Declaration of Helsinki (World Medical, 2013 ) and with data protection and 152 security regulations (Harriss & Atkinson, 2015). 153 Participant characterization 154 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Height and body weight were determined using, respectively, a Harpenden 98.603 stadiometer 155 (Holtain, Crosswell, UK) and a Seca 770 electronic personal scale (Seca, Hamburg, Germany). Body 156 mass composition was assessed using a InBody 770 tetrapolar bioelectrical impedance analyzer 157 (InBody, Cerritos, California, USA) (Brewer et al., 2021). Basal metabolic rate (BMR) was calculated 158 by the analyzer according to Cunnigham equation (BMR (in kcal/day) = 370 + 21.6  fat-free mass (in 159 kg)) (Cunningham, 1991). 160 Acute exercise intervention 161 All participants were asked to perform a maximal incremental voluntary aerobic protocol 162 (Åstrand, 1964 ), which was carried out on an electromagnetically braked Excalibur Sport bicycle 163 ergometer (Lode B.V., Groningen, NL) and was preceded by a 5 -min warm-up of moderate cycling. 164 Throughout the protocol, expired gases were analysed using a Quark CPET breath -by-breath 165 automated gas -analysis system (COSMED, Rome, Italy), which was also used for continuous 166 telemetric determination of heart rate. Initially set at 75 W, the power output was increased by 25-W 167 increments every 3 min, with participants maintaining a sustained cadence of 85 –105 rotations per 168 minute (rpm), until volitional exhaustion was reached (i.e., until participants expressed their inability 169 to continue exercising) or when at least two of the following three criteria were met: (i) no increase in 170 oxygen consumption (VO2) despite workload increase; (ii) respiratory exchange ratio (ratio of volume 171 of CO 2 produced to volume of O 2 consumed) > 1.10; (iii) heart rate above 90% of the estimated 172 maximal value (Beaver et al., 1986; Edvardsen et al., 2014). Ratings of perceived exertion using the 173 Borg CR‐10 scale (Borg, 1982) were self-assessed at each stage and at the end of the protocol. VO2 174 max, i.e., the maximum rate of O 2 consumption attained during physical exertion , was considered as 175 the highest of the mean VO 2 values calculated for the two 30 -s periods that preceded the end of the 176 intervention (Beaver et al., 1986; Edvardsen et al., 2014). Index finger capillary L-lactate levels before 177 and at the end of the exercise intervention were determined using a Lactate Pro2 portable meter 178 (Arkray, Amstelveen, NL). 179 Blood collection and serum preparation and storage 180 At two time points (15 min before the intervention and within 5 min into active recovery), whole 181 blood (10–15 mL) was collected from each participant, immediately transferred to a BD Vacutainer™ 182 SST™ II Advance serum separator tube (Becton Dickinson and Company, Franklin Lakes, NJ, USA) 183 and allowed to clot naturally by leaving it undisturbed, at room temperature. After centrifugation (10 184 min at 1,600 g), the off-the-clot serum was carefully removed and immediately stored at −80 °C, in 185 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint 0.5 mL aliquots, until use. Before use, sera were thawed and, after homogenization, were filtered 186 through 0.2 µm-pore Millex®-GV sterile syringe filters (Merck, Darmstadt, Germany; SLGV033RB). 187 Quantification of glucose, cholesterol, high -density lipoprotein cholesterol , low-density lipoprotein 188 cholesterol, triglycerides and albumin 189 Reagents from BioSystems (Barcelona, Spain) were used for the quantification of glucose 190 (COD 11504), cholesterol (free and esterified; COD 11505), high-density lipoprotein cholesterol 191 (HDL cholesterol; COD 11757), low -density lipoprotein cholesterol (LDL cho lesterol; COD 192 11785), triglycerides (11528) and albumin ( COD 11573) in individual sera, accord ing to the 193 manufacturer’s instructions. A human, serum -based general biochemistry calibrator (COD 194 18044) and human biochemistry control sera (COD 18042 and 18043) , also from BioSystems, 195 were used for calibration and to verify the accuracy of the measurements, respectively. All 196 analytes were assayed in duplicate and the two values were averaged. 197 Quantification of C-reactive protein 198 C-reactive protein (CRP) was quantified by turbidimetry using and assay reagent (COD 199 31321), a calibrator (COD 31113) and a control serum (COD 31213) from Biosystems . All sera 200 were tested in duplicate and the values were then averaged. The procedure was that specified by 201 the manufacturer. 202 Culture of A549 cells 203 All cellular studies were performed in cultures of A549 human lung cancer cells (ATCC 204 CCL-185; RRID:CVCL_0023 ). These cells were routinely grown at 37 °C in a humidified 205 atmosphere of 5% CO 2/95% air, in filter -vented flasks (Orange Scientific, Braine -l’Alleud, 206 Belgium; 5520100)) containing ca. 0.2 mL/cm2 of growth medium (RPMI-1640; Merck; R6504) 207 supplemented with 10% (v/v) heat-inactivated fetal calf serum ( FCS; Thermo Fisher Scientific, 208 Waltham, MA, USA; 10270-106). 209 To investigate the impact of exercise -conditioned human serum on cell proliferation , lag 210 phase and plating efficiency, cells were harvested in serum-free growth medium, centrifuged for 211 5 min at 200 g, to remove traces of FCS, and resuspended in an adequate volume of fresh serum-212 free medium. A small volume (10 µL) of this serum-free cell suspension was subsequently added 213 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint to culture plates already containing pre -warmed (37 °C) growth medium supplemented with human 214 serum, to a final serum concentration of 10% (v/v). 215 To investigate the impact of exercise -conditioned human serum on cell migration and for the 216 determination of half -maximal cytotoxic concentrations (CC 50 values) for cisplatin, cells were 217 harvested in growth medium supplemented with 10% (v/v) heat -inactivated FCS and new cultures 218 were established in this same medium. At ca. 24 h post-seeding, spent growth medium was replaced 219 by growth medium supplemented with 10% (v/v) human serum. 220 Assessment of cell proliferation and lag phase 221 For the assessment of cell proliferation, cultures were prepared in 96 -well plates (Orange 222 Scientific; 4430100), at a seeding density of 5 000 cells/cm 2, in 100 µL of growth medium 223 supplemented with 10% (v/v) human serum. In each independent experiment , a total of six replicate 224 cultures were prepared for each serum tested : three each for the estimation of the number of cells in 225 culture at 24 h and 72 h post-seeding. Cell numbers were estimated using the sulforhodamine B (SRB) 226 assay, essentially as described in (Vichai & Kirtikara, 2006). Briefly, at each of the two time points, 227 cells were fixed using a cold 10% (w/v) trichloroacetic acid (Merck; T6399) solution and were then 228 stained using a 0.05% (w/v) SRB solution (Merck; 230162). After washing four times with a 1% (v/v) 229 acetic acid solution (Merck; A6283), the protein -bound dye was solubilized in 200 µL of a 10 mM 230 Tris base solution, pH 10.5 (Merck; 252859), homogenized and the absorbance of the resulting 231 solutions was read at 550 nm (A550) against a reagent blank (no cells), using a µQuant microplate reader 232 (BioTek Instruments, Winooski, VT, USA). Proliferation rates were then estimated as the fold increase 233 in A550 over the 48-h period between the two measurements. A550 values at 24 h post-seeding was used 234 as a metric for lag phase. 235 Assessment of plating efficiency 236 Plating efficiencies were evaluated by the clonogenic assay (Franken et al., 2006). To this end, 237 cells were plated as a single-cell suspension at a colony-forming density of 40 cells per well in 24-well 238 plates (Orange Scientific; 4430300), in 500 µL of growth medium supplemented with 10% (v/v) 239 human serum. After nine days of incubation, each culture was washed twice with 2 mL of phosphate-240 buffered saline (PBS) and colonies were subsequently fixed and stained for 40 min with 400 μL of an 241 aqueous solution containing 6.0% (v/v) glutaraldehyde (Merck; G5882) and 0.5% (w/v) crystal violet 242 (Merck; 548629). Excess fixing/staining solution was then removed, and cultures were washed several 243 times with water. Finally, wells were photographed, colonies were counted manually, and the area of 244 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint the wells covered by the colonies was measured by ImageJ software version 1.53e (National 245 Institutes of Health, Bethesda, MD, USA). Colony mean size was calculated by dividing this area 246 by the total number of colonies, and plating efficiency was calculated by dividing the number of 247 colonies formed by the number of cells seeded . Each pooled serum was tested in three replicate 248 cultures. 249 Assessment of cell migration 250 In vitro cell migration was estimated employing the wound closure assay (Liang et al., 2007; 251 Cappiello et al. , 2018 ), using IBIDI’s silicone 3 -well culture inserts (IBIDI ®, Gräfelfing , 252 Germany; 80369) to create two artificial 500 µm cell-free gaps in confluent monolayer cultures, 253 corresponding to two technical replicates. These inserts were placed onto the wells of 24 -well 254 plates (Orange Scientific; 4430300), one insert per well and per individual serum tested. Cells 255 were then seeded into the insert wells, at a density of 2.5  105 cells/cm2, in 70 μL of growth 256 medium supplemented with 10% (v/v) FCS. After 22 h of incubation, the inserts were removed 257 to expose two cell -free gaps in each confluent monolayer, growth medium was aspirated, 258 monolayers were washed twice with PBS and three different regions in each gap were imaged 259 (first time point). Pre -warmed (37 °C) growth medium supplemented with 10% (v/v) human 260 serum (500 µL) was then added, cultures were returned to the incubator and, after 18 h, growth 261 medium was once again aspirated, monolayers were washed twice with PBS and the same regions 262 in the gaps were imaged (second time point). All images were captured with a 40 magnification, 263 using an Olympus CKX53 inverted optical microscope equipped with a camera and the EPview™ 264 software (V2.9.6_20201224; Hachioji, Tokyo, Japan). Cell-free areas in the regions imaged were 265 measured using ImageJ and the percentage of gap closure over the 18 -h period was calculated . 266 For each serum tested, the average of the respective six percentage values was used as an estimate 267 of cell migration potential. 268 Determination of CC50 values for cisplatin 269 To determine the concentration of cisplatin that causes 50% of cell death (CC50) in A549 270 cultures, cultures were prepared in 96 -well plates (Orange Scientific), with a seeding density of 271 5 000 cells/cm2, in 100 µL of growth medium supplemented with 10% (v/v) FCS. After a 24 -h 272 incubation, spent growth medium was replaced with pre-warmed (37 °C) growth medium 273 supplemented with 10% (v/v) human serum containing 0–455 μM cisplatin (Merck; 232120) and 274 cultures were further incubated for 72 h. For each pooled serum, the cytotoxicity of each cisplatin 275 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint concentration was tested in three replicate cultures using the SRB assay (as described above) . CC50 276 values were obtained from concentration -response curves generated using a nonlinear regression 277 model in GraphPad Prism 9.0.0. software for Windows (GraphPad Software Inc, Boston, 278 Massachusetts, USA). 279 Semi-quantitative detection of cytokines 280 Semi-quantitative assessment of 80 cytokines was carried out in 0.5 mL samples of pooled sera 281 using the RayBio® C-Series Human Cytokine Antibody Array C5, from RayBiotech (Norcross, GA, 282 USA; AAH-CYT-5-4). Samples of pooled sera were first diluted two -fold with the blocking buffer 283 provided with the kits. Following various blocking, incubation and washing steps, performed 284 according to the manufacturer’s instructions, chemiluminescence was detected using a ChemiDoc MP 285 Imaging System (Bio -Rad, Hercules, CA, USA), after a 30 -second exposure . The signal intensity 286 (average pixel/area) for each antigen-specific antibody spot was then determined using ImageJ. 287 Statistical analysis 288 Data were analyzed using GraphPad Prism 9.0.0. software for Windows. Differences between 289 the two cohorts were assessed for statistical significance using Student's unpaired t-test or the Mann-290 Whitney test, according to the normality of the data, which was verified using the Shapiro -Wilk test. 291 The statistical significance of the differences between the two cohorts in terms of the impact of the 292 acute exercise intervention on serum responses serum-mediated modulation of growth properties and 293 sensitivity to cisplatin cells was assessed using repeated measures two -way ANOVA, assuming 294 sphericity. Whenever the null hypothesis was rejected, Šídák's multiple comparison test was performed 295 (Glantz et al., 2001; Maxwell & Delaney, 2004 ). Differences were considered significant when P < 296 0.05. 297 298 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

299 Cohort characteristics 300 Selected anthropometric and physiological characteristics of the two age -matched male 301 cohorts employed in this study are summarized in Table 1. The most pronounced anthropometric 302 differences between the cohorts were related to body fat content, with non-exercisers exhibiting, 303 on average, ca. 40% higher fat mass ( P ≤ 0.001) and 60% higher visceral fat area ( P = 0.002) 304 than master athletes. Non-exercisers also exhibited ca. 10% lower muscle mass (P = 0.008) and 305 5% higher BMI (P = 0.026). 306 Table 1. Anthropometric and physiological characteristics of the two age -matched male cohorts 307 employed in the present study. 308 Characteristics Non-exercisers (n = 13) Master athletes (n = 13) Statistical significance (P) Age (years) 49 ± 6 53 ± 7 0.155 Height (cm) 172.6 ± 5.7 179.6 ± 9.4 0.031 Body mass index (BMI) (kg/m2) 26.7 ± 2.5 25.3 ± 2.7 0.026 Muscle mass (%) 42.3 ± 2.7 45.7 ± 4.4 0.008 Fat mass (%) 24.7 ± 5.0 17.6 ± 4.8 ≤ 0.001 Visceral fat area (cm2) 89.2 ± 27.3 56.1 ± 16.7 0.002 Basal metabolic rate (kcal/day) 1662.0 ± 152.6 1760.0 ± 185.3 0.168 Blood lactate levels at rest † (mmol dm−3) 1.9 ± 0.6 1.7 ± 0.5 0.531 Blood lactate levels post- intervention # (mmol dm−3) 12.7 ± 2.8 11.1 ± 3.0 0.175 VO2 max (mL/min/kg) 31.7 ± 5.0 39.6 ± 5.6 ≤ 0.001 Duration of the exercise intervention # (min) 13.4 ± 3.3 20.7 ± 4.1 ≤ 0.001 Maximum power output # (W) 170.0 ± 34.8 232.5 ± 35.5 ≤ 0.001 Values represent mean ± standard deviation for the 13 participants of each cohort, except for capillary 309 blood lactate levels, where values from some participants were excluded for technical reasons (one 310 non-exerciser and three master athletes, for at rest levels; one master athlete for post -intervention 311 values). Comparisons between the two cohorts were made using either Student’s unpaired t test (age, 312 height, fat mass, visceral fat, blood lactate levels, VO 2 max, duration of the exercise intervention and 313 maximum power output) or the Mann -Whitney test (BMI, muscle mass and basal metabolic rate), 314 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint according to the normality of data, which was assessed using the Shapiro -Wilk test. Statistical 315 significance was set at P < 0.05 and is identified in bold type. VO 2 max, maximum rate of O 2 316 consumption attained during physical exertion. † Index fingertip capillary L -lactate levels. # The 317 intervention was stopped when the participant reached volitional exhaustion or when at least two of 318 the following three criteria were met: (i) no increase in oxygen consumption (VO 2) despite workload 319 increase; (ii) respiratory exchange ratio > 1.10; (iii) heart rate above 90% of the estimated maximal 320 value. 321 As expected, master athletes exhibited a significantly higher cardiorespiratory fitness than non -322 exercisers (Rogers et al. , 1990 ; Mckendry et al. , 2018 ), as assessed by the maximum rate of O 2 323 consumption attained during physical exertion (VO 2 max) and by the duration of the exercise 324 intervention, whose mean values were, respectively, ca. 25% (P ≤ 0.001) and 50% higher (P ≤ 0.001) 325 than those of non-exercisers. In line with the increase in the duration of the intervention, the maximum 326 power output increased by ca. 35% (P ≤ 0.001). Basal metabolic rate and blood lactate levels (at rest 327 and after the exercise intervention), on the other hand, were not significantly different between the two 328 cohorts. However, the increase in blood lactate levels produced by the exercise intervention differed 329 substantial amongst the different participants of each cohort, with individual increases ranging between 330 5.2-fold and 9.5 -fold, amongst non -exercisers, and between 2.5 -fold and 12.1 -fold amongst master 331 athletes (data not shown). These differences did not correlate with either the duration of the exercise 332 intervention or the maximal power output or VO2 max (Pearson’s determination coefficients for these 333 hypothetical correlations, R2, ranged between 0.004 and 0.126, and the corresponding P values ranged 334 between 0.081 and 0.780). 335 Baseline serum levels of CRP, a biomarker of inflammation (Sproston & Ashworth, 2018), were 336 also determined for each participant. Of the 13 non -exercisers, 12 had normal CRP levels (< 0.5 337 mg/dL), whereas one had a slightly increased level (0.72 mg/dL). All master athletes had normal CRP 338 levels. CRP levels were below the limit of detection of the standard CRP test employed (0.1 mg/dL) 339 in sera from 5 non-exercisers and in sera from 10 master athletes (data not shown). None of the small 340 differences in the cohorts’ serum levels of glucose, lipids, and albumin were statistically significant 341 (Table 2). Also, the effect of the exercise intervention in these levels was equivalent for all analytes 342 tested (all increased by ca. 10% in both cohorts; results not shown). As such, rather than a specific 343 response to acute exercise, the increase may reflect a transient small reduction of plasma volume due 344 to the intervention, as previously reported for both normotensive and untreated hypertensive patients 345 (Hansen, 1968). 346 Table 2. Baseline serum levels of glucose, cholesterol, triglycerides and albumin in the two age -347 matched cohorts employed in the present study. 348 Analyte Non-exercisers Master athletes Statistical significance .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint (n = 13) (n = 13) (P) Glucose (mg/dL) 104.7 ± 24.6 97.6 ± 17.0 0.605 Cholesterol (mg/dL) 212.6 ± 32.4 218.1 ± 29.6 0.654 HDL cholesterol (mg/dL) 49.7 ± 12.2 57.4 ± 7.5 0.065 LDL cholesterol (mg/dL) 126.3 ± 27.4 123.4 ± 22.3 0.797 Triglycerides (mg/dL) Albumin (g/L) 128.6 ± 34.0 50.2 ± 7.4 106.4 ± 40.5 52.1 ± 2.0 0.143 0.373 Each serum was tested twice for the different analytes and the two results were averaged. Values 349 represent mean ± standard deviation for the averages of the 13 participants in each cohort. Comparisons 350 between the two cohorts were made using either unpaired Student’s t-test (cholesterol, HDL 351 cholesterol, LDL cholesterol, triglycerides and albumin) or the Mann -Whitney test (glucose), 352 according to the normality of data, which was assessed using the Shapiro -Wilk test. For all analytes 353 tested, no statistically significant differences were found between the two cohorts (statistical 354 significance was set at P < 0.05). HDL, High-density lipoprotein; LDL, Low-density lipoprotein. 355 Serum conditioning by acute exercise reduced the proliferation rate of A549 cells, whereas 356 conditioning by lifelong exercise had the opposite effect 357 As can be appreciated in Figure 1, acute and lifelong exercise had opposite effects on cell 358 proliferation: whereas serum conditioned by acute exercise decreased cell proliferation by ca. 359 10% (P = 0.062), for non -exercisers, and by ca. 15% (P = 0.003), for master athletes , serum 360 conditioned by lifelong exercise stimulated it by ca. 10% (P = 0.035). The effect of acute exercise 361 was highly consistent in the masters’ cohort, being observed in 11 out of the 13 (85%) individuals, 362 but less consistent in the non -exercisers’ cohort, where it was only observed in 8 out of the 13 363 (62%) individuals. Nonetheless, the decrease in proliferation rate produced by the intervention 364 did not differ, on average, significantly between the two cohorts (P = 0.406). 365 366 A B .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint C Figure 1. Serum conditioning by acute and lifelong exercise had opposite effects on the proliferation 367 rate of A549 human lung cancer cells. ( A) For most of the participants in both cohorts, conditioning 368 by acute exercise reduced sera’s ability to stimulate cell proliferation. ( B) On average, the impact of 369 the acute exercise intervention on sera´s ability to stimulate cell proliferation did not differ 370 significantly (ns) between the two cohorts (right panel). ( C) On average, cells grown in medium 371 supplemented with baseline masters’ serum exhibited higher proliferation rates than those grown in 372 medium supplemented with baseline non-exercisers’ serum. All cultures were established and grown 373 in medium supplemented with 10% (v/v) human serum. The number of cells in culture at 24 h and 48 374 h post-seeding, used to calculate proliferation rates, was estimated using the sulforhodamine B (SRB) 375 assay. Dots and squares represent individual values for the different sera and are the means of three 376 independent experiments. In each of these experiments, all conditions (sera and time points) were 377 tested in three replicate cultures. Each set of connected dots represents sera collected from the same 378 participant before (left; Baseline) and after (right; Post -intervention) the acute exercise intervention. 379 Large horizontal bars and associated smaller horizontal (error) bars represent means ± standard 380 deviation for the 13 participants of each cohort. The statistical significance of the effect of lifelong 381 exercise was assessed using Student's unpaired t-test. All other comparisons were made by repeated 382 measures two-way ANOVA, assuming sphericity. Whenever the null hypothesis was rejected, Šídák's 383 multiple comparison test was performed. Statistical significance was set at P < 0.05 and statistically 384 significant differences between the indicated groups are shown by *, for P ˂ 0.05, and by **, for P < 385 0.01. 386 Serum conditioning by acute exercise increased the lag phase of A549 cells, whereas conditioning by 387 lifelong exercise had the opposite effect 388 When assessing proliferation rates, we observed, for both cohorts, that the number of cells in 389 culture 24 h post-seeding was, on average, ca. 20% higher in cultures stimulated by post-intervention 390 serum than by the respective baseline serum (P ˂ 0.0001 for both cohorts; Figure 2 A,B). Of note, this 391 was highly consistent, being observed in 12 out of 13 (92%) non -exercisers and in all (100%) master 392 athletes. These differences in cell number cannot be explained by different proliferation rates, as our 393 proliferation data point in the opposite direction, i.e., cultures with the higher number of cells 24 h 394 post-seeding had lower proliferation rates (Figure 1). Also, microscopic observation of the cultures did 395 not reveal any differences in the number of dead or unattached cells. Therefore, cells exposed to sera 396 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint conditioned by acute exercise took less time to attach to the substrate, spread and/or resume 397 proliferation than those exposed to the corresponding baseline sera. Surprisingly, conditioning of sera 398 by lifelong exercise had the opposite effect, resulting in ca. 10% decrease in the number of cells in 399 culture 24 h post-seeding (P = 0.281; Figure 2C) suggesting and increased lag phase. 400 A B C Figure 2. Serum conditioning by lifelong and acute exercise had opposite effects on the number of 401 A549 human lung cancer cells in culture 24 h post -seeding. (A) For both cohorts, cultures grown in 402 medium supplemented with serum conditioned by the acute exercise intervention (post -intervention 403 serum) contained, on average, ca. 20% more cells 24 h post -seeding than those grown in medium 404 supplemented with baseline serum from the same participant, suggesting that the serum changes 405 produced by the intervention shortened the cells’ lag phase. This increase was observed in 12 out of 406 13 (92%) non-exercisers and in all (100%) master athletes. (B) No statistically significant differences 407 (ns) were obtained between the two cohorts in terms of the impact of acute exercise serum conditioning 408 on the number of cells in culture 24 h post -seeding. ( C) On average, cultures grown in medium 409 supplemented with baseline masters’ serum contained ca. 10% less cells 24 h post-seeding than those 410 grown in medium supplemented with baseline non -exercisers’ serum. All cultures were established 411 and grown in medium supplemented with 10% (v/v) human serum. The total amount of cell protein, 412 assessed using the sulforhodamine B (SRB) assay, was used as an estimate of the number of cells in 413 culture. Dots and squares represent individual values for the different sera and are the means of three 414 independent experiments. In each of these experiments, all sera were tested in three replicate cultures 415 and the values were averaged. Each set of connected dots represents baseline (left) and post -416 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint intervention (right) sera from the same participant. Large horizontal bars and associated smaller 417 horizontal (error) bars represent means ± standard deviation for the 13 participants of each cohort. The 418 statistical significance of the effect of lifelong exercise was assessed using the Mann-Whitney test. All 419 other comparisons were made by repeated measures two -way ANOVA, assuming sphericity. 420 Whenever the null hypothesis was rejected, Šídák's multiple comparison test was performed. Statistical 421 significance was set at P < 0.05 and statistically significant differences between the indicated groups 422 are shown by ****, for P ˂ 0.0001. A550, Absorbance at 550 nm. 423 Neither acute nor lifelong exercise serum conditioning affected the reproductive potential or the 424 migratory capacity of A549 cells 425 We next investigated the impact of exercise on the ability of cells to form colonies when seeded at 426 very low densities, i.e., on their plating efficiencies, used here as a metric for reproductive potential. 427 Our results show that neither acute nor lifelong exercise had a significative impact on this parameter 428 (Figure 3A–C; P = 0.692 and P = 0.940, for the effect acute exercise on non -exercisers and master 429 athletes, respectively, and P = 0.396 for the effect of lifelong exercise) . Regarding the very large 430 changes in colony size observed ( Figure 3D–F), they are likely the result of the alterations in 431 proliferation rates discussed above (Figure 1). It must be noted that small differences in proliferation 432 rate, such as those observed in our study, can translate into markedly different cell numbers at the end 433 of 9 days (the duration of the assay employed). 434 A B C D .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint E F G Figure 3. Neither acute nor lifelong exercise serum conditioning affected the reproductive potential of 435 A549 human lung cancer cells, as assessed by their plating efficiencies. Pooled sera were used for the 436 four conditions tested (baseline non -exercisers; baseline master athletes; post -intervention non -437 exercisers; post -intervention master athletes), each containing equal volumes of serum from 13 438 participants. (A–C) Plating efficiencies, expressed as percentage values, were calculated by dividing 439 the number of colonies formed by the number of cells seeded, whereas (D–F) colony mean sizes were 440 calculated dividing the area covered by the colonies (measured using ImageJ) by the total number of 441 colonies. Each set of three dots and squares represents the values obtained in three independent 442 experiments. In each of these experiments, all pooled sera were tested in three replicate cultures and 443 the results averaged. Large horizontal bars and associated smaller horizontal (error) bars represent 444 means ± standard deviation for the three independent experiments. The statistical significance of the 445 effect of lifelong exercise was assessed using Student's unpaired t-test. All other comparisons were 446 made by repeated measures two -way ANOVA, assuming sphericity. Whenever the null hypothesis 447 was rejected, Šídák's multiple comparison test was performed. Statistical significance was set at P < 448 0.05 and statistically significant differences between the indicated groups are shown by **, for P ˂ 449 0.01, and ***, for P ˂ 0.001. ns, No statistically significant differences were obtained (i) in plating 450 efficiency between baseline and corresponding post -intervention pooled sera in the two cohorts and 451 between baseline sera from the two cohorts; (ii) in the mean size of the colonies between baseline and 452 post-intervention sera in non -exercisers. (G) Representative photographs of the colonies that formed 453 after 9 days of incubation when cells were plated using a single -cell suspension at a colony -forming 454 density of 40 cells per well in 24 -well plates, in 500 µL of growth medium supplemented with 10% 455 (v/v) of pooled human serum. 456 The results of our cell migration assessments are summarized in Figure 4. As can be 457 appreciated, neither acute nor lifelong exercise conditioning altered serum’s ability to promote 458 cell migration to a statistically significant extent (Figures 4A–C; P = 0.998 and P = 0.555, for the 459 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint effect of acute exercise on non-exercisers and master athletes, respectively; P = 0.448, for the eff ect 460 of lifelong exercise). It must be noted, though, that cells were exposed to human sera for a relatively 461 short period (18 h). Also, due to the limited volume of human serum available, the results are from a 462 single experiment (in which all 52 sera were tested in duplicate). 463 A B C D Figure 4. Neither acute nor lifelong exercise serum conditioning affected the in vitro migratory 464 capacity of A549 human lung cancer cells. (A–C) To assess in vitro migratory capacity, two artificial 465 500 µm cell-free gaps were created in confluent monolayer cultures, using 3 -well culture inserts that 466 were placed onto the wells of 24 -well plates, one insert per well and per serum tested. Cultures were 467 then exposed to medium supplemented with 10% (v/v) human serum. For each gap, three different 468 regions were imaged immediately after its creation and after an 18-hour incubation. Cell-free areas in 469 these images were then measured using ImageJ and the percentage of gap closure over this period was 470 calculated. For each serum, the average of the corresponding six values was then used as an estimate 471 of cell migration potential. Dots and squares represent individual values for the different sera. Each set 472 of connected dots represents baseline (left) and post-intervention (right) sera from the same participant. 473 Large horizontal bars and associated smaller horizontal (error) bars represent means ± standard 474 deviation for the 13 participants of each cohort. The statistical significance of the effect of lifelong 475 exercise was assessed using the Mann -Whitney test. All other comparisons were made by repeated 476 measures two-way ANOVA, assuming sphericity. Statistical significance was set at P < 0.05 and the 477 null hypothesis was never rejected (ns). Results are from a single independent experiment in which 478 each serum was tested in two artificial gaps. ( D) Representative micrographs (40  magnification) of 479 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint artificial gaps for each of the four conditions tested, taken immediately after the creation of the gap 480 and after an 18 -h incubation in growth medium supplemented with 10% (v/v) human serum. 481 Micrographs were captured using an Olympus CKX53 inverted optical microscope equipped with a 482 camera and the EPview™ software (V2.9.6_20201224; Hachioji, Tokyo, Japan). 483 Lifelong exercise increased the sensitivity of A549 cells to cisplatin 484 As can be appreciated in Figure 5, sensitivity to cisplatin was much higher (as assessed by 485 the lower CC50 value) when cells were incubated in the presence of baseline serum from master 486 athletes, than baseline serum from non-exercisers (P = 0.04). On the contrary, acute exercise had 487 no significant impact on this sensitivity (P = 0.999 and P = 0.712, for non-exercisers and master 488 athletes, respectively). 489 A B Figure 5. Lifelong exercise more than doubled the sensitivity of A549 human lung cancer cells to 490 cisplatin. ( A) Concentration -response curves depicting the cytotoxicity of cisplatin against cells 491 stimulated with pooled sera (10% (v/v)) from non -exercisers and master athletes, obtained at rest 492 (Baseline) and after an acute exercise intervention (Post -intervention). (B) CC50 values for cisplatin 493 for cultures incubated with the four different pooled sera. Cell viability was estimated at 72 h post -494 cisplatin addition using the sulforhodamine B (SRB) assay. For each pooled serum, the impact of the 495 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint different cisplatin concentrations (0 –455 μM) on cell viability was tested in three independent 496 experiments, each with three replicate cultures per concentration, and expressed as percentage of the 497 control value (0 µM cisplatin). CC 50 values were calculated from concentration -response curves 498 generated using a nonlinear regression model in GraphPad Prism. Dots/diamonds and corresponding 499 error bars represent, respectively, means and standard deviation for the three independent experiments. 500 Comparisons between the indicated groups were made by both (A) the sum-of-squares using the extra 501 sum-of-squares F test; (B) Student's unpaired t-test (effect of lifelong exercise) and repeated measures 502 two-way ANOVA, assuming sphericity (effect of acute exercise). Whenever the null hypothesis was 503 rejected, Šídák's multiple comparison test was performed. Statistical significance was set at P < 0.05 504 and statistically significant differences between the indicated groups are shown by *, for P ˂ 0.05, and 505 by **, for P < 0.01. CC50, Half-maximal cytotoxic concentration. 506 Both acute and lifelong exercise modulated serum cytokine levels 507 Acute exercise increased the levels of most cytokines (ca. 50 in both cohorts) and these increases 508 tended to be substantial (up to 7-fold). Unlike in non-exercisers' serum, where there was a very large 509 increase in IL -1, IL-2 and IL -15, among others, and a very large decrease in IL -6, among others, 510 following acute exercise, no effect was observed in the levels of these cytokines in master athletes' 511 serum. Interestingly, for some cytokines, such as GRO (//), IL-8, IL-10 and MIP-1, acute exercise 512 had opposite effects in the two cohorts, reducing their levels in non-exercisers, but augmenting them 513 in master athletes (Figure 6A). 514 Regarding lifelong exercise, it had no effect on ca. 40% of the cytokines tested and when it did 515 have an effect it was mostly a decrease: 32 cytokines saw their signal intensity decrease . The 14 516 cytokines increased, most notably MCP2, SDF1, and TGF-1 (Figure 6B), displayed very weak signals 517 near the limit of detection in both cohorts and their increase may thus not represent true physiological 518 changes. On the contrary, many of the cytokines whose signals were significantly reduced in the serum 519 of master athletes compared to the serum of non-exercisers, namely GM-CSF, GRO (//), GRO-, 520 IL-8 (CXCL8), IL-10, MIP-1 (CCL4), NT-3 and osteopontin, were significantly expressed in both 521 cohorts. 522 523 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint A B .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Figure 6. The effects of exercise on serum cytokine levels depended on cytokine, duration of the 524 exercise (acute versus lifelong) and previous levels of physical activity (non-exercisers versus master 525 athletes). The semi-quantitative assessment of the 80 cytokines was carried out in pooled sera, each 526 containing equal volumes of serum from 13 participants, using an antibody-based microarray. For each 527 antigen-specific antibody spot, chemiluminescence signal intensity (average pixel/area) was 528 determined using ImageJ. Two independent determinations were carried out and the results were 529 averaged (A,B; Left) Cytokine levels are presented in decreasing order of signal intensity in the pooled 530 non-exercisers’ serum. (A,B; Right) Cytokines are presented in order of fold change in signal intensity 531 produced by the acute exercise intervention in the non-exercisers’ serum. 532 533 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

534 Over 30% of the worldwide population does not attain the levels of physical activity for 535 health and well -being recommended by the WHO (Bull et al., 2020), (Elgaddal et al., 2022). 536 Establishing a causal relationship between increased levels of physical activity and lower cancer 537 risk and mortality would constitute a strong inducement for inactive people to significantly reduce 538 the amount of time spent in sedentary occupations. Towards this goal, several research groups 539 exposed cultures of human cancer cell lines to human serum obtained before and after an acute 540 bout of exercise to determine whether the transient systemic changes induced directly impacted 541 their behavior (Rundqvist et al., 2013; Dethlefsen et al., 2016; Dethlefsen et al., 2017; Kurgan et 542 al., 2017; De Santi et al., 2019; Devin et al., 2019; Baldelli et al., 2020; Hwang et al., 2020; 543 Orange et al., 2022; Kim et al., 2023). These studies focused mostly on cell proliferation, which 544 was often assessed by the number of cells in culture after a specified period of incubation. Other 545 parameters were also investigated, but not as consistently, namely apoptosis levels (Leung et al., 546 2004; Rundqvist et al., 2013; Devin et al., 2019) and the cells’ ability to form colonies when 547 seeded at very low densities (Kurgan et al., 2017) or under conditions that prevent cell attachment 548 (De Santi et al., 2019; Baldelli et al., 2020). Despite the scarcity of data and significant inter-549 study heterogeneity , meta -analysis of the results indicates that serum conditioning by acute 550 exercise significantly reduced cancer cell proliferation, with a large effect size that increased with 551 the intensity of the exercise (Soares et al. , 2021 ). Of note, for all cell lines/cohorts reported 552 (Rundqvist et al., 2013; Dethlefsen et al., 2016; Dethlefsen et al., 2017; Kurgan et al., 2017; 553 Devin et al., 2019; Baldelli et al., 2020; Hwang et al., 2020; Orange et al., 2022; Kim et al., 554 2023), only in one single case (PC3 cells and a young age cohort) did acute exercise failed to 555 decrease cell proliferation (Hwang et al., 2020). Significantly, in a study that employed both 556 normal (MRC5) and cancer (A549) lung cell lines , the proliferation of normal cells was not 557 affected, while that of cancer cells was (Kurgan et al., 2017).The effects of chronic exercise on 558 the properties of cultured human cancer cells, on the other hand, have not yet been the subject of 559 meta-analysis, due to the reduced number of studies (Barnard et al., 2003; Leung et al., 2004; 560 Dethlefsen et al., 2016; Baldelli et al., 2020; Hwang et al., 2020; Kim et al., 2022). Also, the 561 duration of the exercise intervention in most of these studies was relatively short (1 to 6 months) 562 (Dethlefsen et al., 2016; Baldelli et al., 2020; Hwang et al., 2020; Kim et al., 2022), with only 563 two studies exploring the effects of exercise training over 10 or more years (Barnard et al., 2003; 564 Leung et al., 2004). Notwithstanding the scarcity of data, the results from these studies suggest 565 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint that the impact of chronic exercise on the ability of sera to stimulate cancer cell proliferation differs 566 from that of sera conditioned by acute exercise (see below). 567 The aim of the present investigation was three-fold. First, to strengthen the evidence regarding 568 the impact of exercise conditioning on the proliferation of cultured human cancer cells . Second, to 569 expand this line of investigation to cell migration, a cell property also frequently used as a marker of 570 transformation degree and, ultimately, tumorigenic potential, which was not previously investigated 571 in this type of study. Compromising or losing any of these markers might ultimately halt or reve rse 572 cancer progression. Last, but not least, to investigate, for the first time, whether the decreased cancer 573 mortality associated with exercise might be partially explained by an increased sensitivity of cancer 574 cells to chemotherap y agents. Our study employed two cohorts of age -matched individuals that 575 differed significantly in terms of levels of physical activity, specifically individuals who did not meet 576 the WHO guidelines on physical activity and master athletes (as a model of lifelong exercise) . The 577 simultaneous use of two distinct cohorts allowed to gain insight into the dependency of the cellular 578 effects induced by acute exercise on cohort characteristics. 579 All cell studies were carried out on cultures of A549 cells, which were used as an in vitro model 580 of human lung cancer, for which epidemiological evidence linking physical activity with reduced 581 cancer risk is strong (Moore et al., 2016). Cellular studies are prone to a high degree of variability, due 582 namely to intra-cell-line heterogeneity. To increase the robustness of our data, multiple biologically 583 independent experiments were carried out , and each serum was tested on at least two technical 584 replicates on each of these experiments. 585 For our analysis of the impact of exercise on proliferation rate, we employed the SRB assay 586 (Vichai & Kirtikara, 2006). This assay quantifies total cellular protein, from which cell numbers and, 587 ultimately, cell proliferation can be estimated. The number of cells in culture at any given time after 588 seeding depends on parameters other than proliferation rates, namely levels of cell death and the time 589 cells take to adhere to the substrate, spread and resume growth and division (i.e., the duration of the 590 cells’ lag phase). To omit the lag phase from our estimation of proliferation rates, cell numbers were 591 assessed at two time points (24 h and 72 h post-seeding), rather than just at the end of the experiment, 592 as was frequently the case in similar studies. Proliferation was then estimated as the fold increase in 593 the number of cells over the 48 -h period. Also, instead of establishing cultures in growth medium 594 supplemented with FCS and only exposing them to human serum 24 h post-seeding, as was also often 595 the case, our cultures were established in growth medium supplemented with human serum, i.e., they 596 were exposed to human serum throughout the experiment. This protocol not only yielded more reliable 597 proliferation data, but also enabled us to gain insight into the influence of exercise conditioning on the 598 lag phase. 599 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Our results support previous findings that acute exercise decreases cancer cell proliferation 600 (Figure 1A,B). Lifelong exercise, on the other hand, increased cell proliferation (Figure 1C). This 601 contrasts with the reports that 10+ years of training decreased the proliferation of LNCaP prostate 602 cancer cells, as determined by end-point analysis, and did not alter the proliferation of a LNCaP-603 derived cell line with nonfunctional p53 (LN -56 cells), suggesting that the effects on cell 604 proliferation depend on the cells’ characteristics, namely their p53 status (Barnard et al., 2003; 605 Leung et al., 2004). It also contrast with the results of three of the studies involving short-term 606 exercise programs (1 –6 months), where no effect on proliferation wa s observed (Dethlefsen et 607 al., 2016; Devin et al., 2019; Baldelli et al., 2020; Kim et al., 2022). Significantly, all these three 608 studies also assessed acute exercise, which decreased cell proliferation. This contrasting pattern 609 between the effects of acute and chronic exercise led the authors to hypothesize that the putative 610 beneficial effects of long-term exercise training on cancer incidence, recurrence and survival 611

from the cumulative effects of transient serum changes in response to each repeated 612 exercise bout, rather than from permanent serum alterations produced by exercise training 613 (Dethlefsen et al., 2016; Devin et al., 2019; Baldelli et al., 2020). However, it must be stressed 614 that the exercise programs lasted 9 weeks at the most, which might not have been sufficient to 615 produce permanent serum changes to factors affecting cell proliferation. 616 A contrasting pattern was also found in terms of the impact of the two different types of 617 exercise on the lag phase, with acute exercise decreasing it (Figure 2A,B), and lifelong exercise 618 extending it (Figure 2C). As the present study was the first to investigate the effects of exercise 619 on this property, it is not possible to tell whether this effect is specific to the A549 cell line and/or 620 the cohorts used, or a more general one. It is known that interactions between cells and the 621 extracellular matrix (ECM) modulate cell survival and proliferation (Frantz et al. , 2010 ). 622 However, considering the limited evidence available, it would not be advisable to speculate on 623 possible consequences of such effect on the cells’ tumorigenic potential. In addition, it must be 624 acknowledged that, when it comes to cancer, adhesion to the ECM can be seen as a double-edged 625 sword. In fact, in order to metastasize, cancer cells must first detach from neighboring cells and 626 the ECM and then resist anoikis during their migration to other parts of the body. But they then 627 need to reattach and resume proliferation, to form tumors in new locations (Welch & Hurst, 2019; 628 Shaw et al., 2025). In future studies, it would be important to investigate the ability of exercise 629 to alter the expression of adhesion molecules, such as cadherins and integrins, and the 630 physicochemical properties of the ECM. 631 Neither acute (Figure 3A,B) nor lifelong exercise (Figure 3C) produced significant changes 632 in plating efficienc y, indicating that the reproductive potential was unchanged. This finding 633 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint contrasts with a previously reported significant decrease in the number of colonies formed by A549 634 and two other cell lines when sera were conditioned by acute exercise (Kurgan et al., 2017). Once 635 again, these distinct outcomes might be due to differences in study design, namely in the criteria used 636 for counting colonies. 637 Together with cell –cell adhesion and cell adhesion to the substrate, the migratory capacity is 638 often used as a metric for metastatic potential (Mehanna et al. , 2025 ). In our study, neither acute 639 (Figure 4A,B) nor lifelong exercise (Figure 4C) altered the migratory capacity of A549 cells. However, 640 cells were exposed to human serum for a relatively short period, and it cannot be excluded that a longer 641 exposure could have produced a different outcome. 642 A most striking result of our study was the strong impact that lifelong exercise had on the 643 sensitivity to cisplatin, which more than doubled (Figure 5). Thus, one of the mechanisms by which 644 exercise might decrease cancer mortality, besides improved immunosurveillance (Bigley et al., 2014; 645 Moro-García et al., 2014), is by increasing the cytotoxicity of chemotherap y agents against cancer 646 cells, thus improving therapeutic efficacy for patients. This is the first report of this type of effect and, 647 due to its significance, this approach should undoubtedly be further pursued. 648 Aiming to shed light on the molecular mechanisms behind the distinct impacts of acute and 649 lifelong serum conditioning on the behavior of cancer cells, we interrogated cytokine levels in sera of 650 non-exercisers and master athletes at baseline and after acute exercise (Figure 6). It is well known that 651 skeletal muscle acts as an endocrine organ that produces and secretes hundreds of cytokines and other 652 signaling peptides, namely in response to exercise. Collectively known as myokines, these molecules 653 mediate communication within skeletal muscle and between skeletal muscle and other organs, playing 654 a wide array of functions (Severinsen & Pedersen, 2020). Myokines such as decorin, irisin, oncostatin 655 M and SPARC have been implicated in the preventive and therapeutic effects of exercise and can 656 directly affect cancer cell behavior by several means, namely by inhibiting proliferation, promoting 657 apoptosis and inhibiting epithelial to mesenchymal cells transition, ultimately limiting invasion and 658 metastasis. Myokines can also reduce cancer risk by indirect means, namely by inhibiting fat 659 accumulation ( e.g., IL -6, IL -15, irisin and SPARC), reversing insulin resistance, reducing chronic 660 inflammation and through modulation of the immune system (Kim et al., 2021; Huang et al., 2022). 661 Direct comparisons of the present work with other studies or between our two cohorts are not 662 straightforward, as it has been found that exercise -mediated cytokine secretion depends on the 663 characteristics of the exercise, such as type, intensity and duration. For instance, in the case of some 664 cytokines, a significant response was only observed with more than 30 minutes of sustained moderate-665 to-high-intensity exercise involving major muscle groups (Piccirillo, 2019; Kim et al., 2021). Also, 666 cytokine serum levels are notoriously small (in the picomolar or femtomolar range) (Anderson & 667 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Anderson, 2002). Accordingly, chemiluminescent signals for some cytokines were very weak. 668 Changes in signal intensity produced by exercise below 20% were thus regarded as not significant. 669 Also, increases/decreases in signal intensity above 20% must be interpreted with care in the case 670 of cytokines whose signals were very weak in the two conditions under comparison, as small 671 changes in signal intensity translate into marked fold changes, which do not necessarily have 672 biological relevance. Nonetheless, some comments may be tentatively made. Namely that the 673 effects on the serum levels of many cytokines were dependent on the type of exercise ( acute 674 versus lifelong) and cohort, as was the case with the observed effects on cell growth 675 characteristics. Also, most of the cytokines that were affected by acute exercise had their levels 676 increased, whereas the opposite was true for lifelong exercise. Of note, despite the changes it 677 induced, lifelong exercise did not substantially alter the pattern of signal intensity: cytokines 678 present at high levels in non -exercisers were also highly expressed in master athletes, and the 679 same was true for minimally expressed cytokines. Globally speaking, the increases produced by 680 acute exercise were more pronounced in non -exercisers than in the master athletes. For some 681 cytokines, such as IL -8, GRO ( //), and MIP -1, chemokines responsible for recruiting 682 neutrophils and monocytes to sites of inflammation (Fujiwara et al., 2002; Parekh et al., 2019), 683 and IL-10, responsible for the anti-inflammatory effect of exercise (Gleeson et al., 2011), acute 684 exercise had opposite effects in the two cohorts, reducing their serum levels in non -exercisers, 685 but augmenting them in master athletes. Of note, serum levels of several cytokines linked to 686 the senescence-associated secretory phenotype (Li et al., 2023) were lower in master athletes than 687 in non-exercisers, namely those of eotaxin -3 (CCL26), GM -CSF, GRO, IL -8, MIP-1β (CCL4) 688 and MIP-1. Indeed, upon acquisition of this phenotype, fibroblasts become proinflammatory 689 cells with the ability to promote tumor progression (Coppé et al., 2010). 690 691 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

692 Our study is the first to address the impact of exercise on the susceptibility of cultured cancer 693 cells to a chemotherapy drug (cisplatin), as well as on a series of properties that characterize the growth 694 pattern of cultured human cancer cells, namely migratory capacity and cell adhesion to the substrate, 695 resumption of cell growth and division after detachment. Other strengths of our study include the 696 parallel investigation of the effects of both transient and permanent serum responses to, respectively, 697 acute and lifelong exercise. Of note, we are the first to examine master athletes in this type of study 698 and to assess the effects of acute exercise on two cohorts differing significantly in their baseline 699 exercise levels in parallel. 700 In summary, we found, for both cohorts, that acute exercise decreased the proliferation rate of 701 A549 cells, as previously reported by several groups for this and other cell lines. Interestingly, while 702 acute exercise decreased the proliferation rate, it shortened the time cells took to adhere to the substrate, 703 spread and resume proliferation. L ifelong exercise had the opposite effects: it increased the 704 proliferation rate yet increased the lag phase. Regarding plating efficiency and migratory capacity, no 705 effect could be detected. A different outcome for migratory capacity might have been observed if the 706 duration of the exposure to human serum had been increased. Strikingly, lifelong exercise more than 707 doubled the sensitivity of cancer cells to cisplatin, which may explain, together with other mechanisms 708 such as enhanced immunosurveillance , the lower cancer mortality rates found among those who 709 exercise regularly. Moreover, serum cytokine patterns induced by lifelong exercise and acute exercise 710 contrasted sharply, potentially contributing to the different impacts that these two types of exercise 711 had on cancer cell properties. 712 713 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint

714 Adam B, Modenese A & Loney T. (2023). Editorial: Occupation and cancer: new insights into burden, 715 risk factors, and prevention. Frontiers in Public Health 11, 1343952. 716 717 Anderson NL & Anderson NG. (2002). The Human Plasma Proteome: History, Character, And 718 Diagnostic Prospects*. Molecular & Cellular Proteomics 1, 845-867. 719 720 Åstrand P -O. (1964). Work tests with the bicycle ergometer. Gymnastik- och Idrottshögskolan 721 Stockholm, Sweden. 722 723 Avgerinos KI, Spyrou N, Mantzoros CS & Dalamaga M. (2019). Obesity and cancer risk: Emerging 724 biological mechanisms and perspectives. Metabolism: Clinical and Experimental 92, 121-135. 725 726 Baldelli G, De Santi M, Gervasi M, Annibalini G, Sisti D, Højman P, Sestili P, Stocchi V, Barbieri E 727 & Brandi G. (2020). The effects of human sera conditioned by high-intensity exercise sessions 728 and training on the tumorigenic potential of cancer cells. Clinical and Translational Oncology 729 23, 22-34. 730 731 Barnard RJ, Ngo TH, Leung PS, Aronson WJ & Golding LA. (2003). A low-fat diet and/or strenuous 732 exercise alters the igf axis in vivo and reduces prostate tumor cell growth in vitro. The Prostate 733 56, 201-206. 734 735 Beaver WL, Wasserman K & Whipp BJ. (1986). A new method for detecting anaerobic threshold by 736 gas exchange. Journal of Applied Physiology 60, 2020-2027. 737 738 Bigley AB, Rezvani K, Chew C, Sekine T, Pistillo M, Crucian B, Bollard CM & Simpson RJ. (2014). 739 Acute exercise preferentially redeploys NK -cells with a highly -differentiated phenotype and 740 augments cytotoxicity against lymphoma and multiple myeloma target cells. Brain, Behavior, 741 and Immunity 39, 160-171. 742 743 Borg GA. (1982). Psychophysical bases of perceived exertion. Medicine and Science in Sports and 744 Exercise 14, 377-381. 745 746 Brewer GJ, Blue MNM, Hirsch KR, Saylor HE, Gould LM, Nelson AG & Smith -Ryan AE. (2021). 747 Validation of inbody 770 bioelectrical impedance analysis compared to a four -compartment 748 model criterion in young adults. Clinical Physiology and Functional Imaging 41, 317-325. 749 750 Brown MJ, Morris MA & Akam EC. (2021). An exploration of the role of exercise in modulating 751 breast cancer progression in vitro: A systematic review and meta-analysis. Cell Physiology 320, 752 C253-C263. 753 754 Bull FC, Al-Ansari SS, Biddle S, Borodulin K & Buman MP. (2020). World health organization 2020 755 guidelines on physical activity and sedentary behaviour. British Journal of Sports Medicine 54, 756 1451-1462. 757 758 Cappiello F, Casciaro B & Mangoni ML. (2018). A novel in vitro wound healing assay to evaluate cell 759 migration. Journal of Visualized Experiments. 760 761 Coppé JP, Desprez PY, Krtolica A & Campisi J. (2010). The senescence -associated secretory 762 phenotype: The dark side of tumor suppression. Annual Review of Pathology 5, 99-118. 763 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint 764 Courneya KS, Vardy JL, O'Callaghan CJ, Gill S, Friedenreich CM, Wong RKS, Dhillon HM, Coyle 765 V, Chua NS, Jonker DJ, Beale PJ, Haider K, Tang PA, Bonaventura T, Wong R, Lim HJ, Burge 766 ME, Hubay S, Sanatani M, Campbell KL, Arthuso FZ, Turner J, Meyer RM, Brundage M, 767 O'Brien P, Tu D & Booth CM. (2025). Structured Exercise after Adjuvant Chemotherapy for 768 Colon Cancer. The New England Journal of Medicine 393, 13-25. 769 770 Cunningham JJ. (1991). Body composition as a determinant of energy expenditure: A synthetic review 771 and a proposed general prediction equation. The American Journal of Clinical Nutrition 54, 772 963-969. 773 774 De Santi M, Baldelli G, Lucertini F, Natalucci V, Brandi G & Barbieri E. (2019). A dataset on the 775 effect of exercise-conditioned human sera in three-dimensional breast cancer cell culture. Data 776 In Brief 27, 104704. 777 778 Dethlefsen C, Hansen LS, Lillelund C, Andersen C, Gehl J, Christensen JF, Pedersen BK & Hojman 779 P. (2017). Exercise-induced catecholamines activate the hippo tumor suppressor pathway to 780 reduce risks of breast cancer development. Cancer Research 77, 4894-4904. 781 782 Dethlefsen C, Lillelund C, Midtgaard J, Andersen C, Pedersen BK, Christensen JF & Hojman P. (2016). 783 Exercise regulates breast cancer cell viability: Systemic training adaptations versus acute 784 exercise responses. Breast Cancer Res Treat 159, 469-479. 785 786 Devin JL, Hill MM, Mourtzakis M, Quadrilatero J, Jenkins DG & Skinner TL. (2019). Acute high 787 intensity interval exercise reduces colon cancer cell growth. J Physiol 597, 2177-2184. 788 789 Edvardsen E, Hem E & Anderssen SA. (2014). End criteria for reaching maximal oxygen uptake must 790 be strict and adjusted to sex and age: A cross-sectional study. PloS one 9, e85276. 791 792 Elgaddal N, Kramarow EA & Reuben C. (2022). Physical activity among adults aged 18 and over: 793 United states, 2020. NCHS Data Brief, 1-8. 794 795 Franken NAP, Rodermond HM, Stap J, Haveman J & Van Bree C. (2006). Clonogenic assay of cells 796 in vitro. Nature Protocols 1, 2315-2319. 797 798 Frantz C, Stewart KM & Weaver VM. (2010). The extracellular matrix at a glance. Journal of Cell 799 Science 123, 4195-4200. 800 801 Friedenreich CM, Neilson HK & Lynch BM. (2010). State of the epidemiological evidence on physical 802 activity and cancer prevention. European Journal of Cancer 46, 2593-2604. 803 804 Fujiwara K, Matsukawa A, Ohkawara S, Takagi K & Yoshinaga M. (2002). Functional distinction 805 between CXC chemokines, interleukin -8 (IL -8), and growth related oncogene (GRO)α in 806 neutrophil infiltration. Laboratory Investigation 82, 15-23. 807 808 Garcia L & Pearce M. (2023). Non -occupational physical activity and risk of cardiovascular disease, 809 cancer and mortality outcomes: A dose -response meta-analysis of large prospective studies. 810 British Journal of Sports Medicine 57, 979-989. 811 812 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Glantz SA, Slinker BK & Neilands TB. (2001). Primer of applied regression & analysis of variance, 813 ed, vol. 654. McGraw-Hill, Inc., New York. 814 815 Gleeson M, Bishop NC, Stensel DJ, Lindley MR, Mastana SS & Nimmo MA. (2011). The anti -816 inflammatory effects of exercise: Mechanisms and implications for the prevention and 817 treatment of disease. Nature Reviews Immunology 11, 607-615. 818 819 Hansen J. (1968). The effect of short -term exercise on plasma volume and blood pressure in 820 guanethidine-treated hypertensives. Acta Medica Scandinavica 183, 553-558. 821 822 Harriss DJ & Atkinson G. (2015). Ethical standards in sport and exercise science research: 2016 update. 823 International Journal of Sports Medicine 36, 1121-1124. 824 825 Huang Q, Wu M, Wu X, Zhang Y & Xia Y. (2022). Muscle-to-tumor crosstalk: The effect of exercise-826 induced myokine on cancer progression. Biochimica et Biophysica Acta Reviews on Cancer 827 1877, 188761. 828 829 Hwang JH, Mcgovern J, Minett GM, Della Gatta PA, Roberts L, Harris JM, Thompson EW, Parker 830 TJ, Peake JM & Neubauer O. (2020). Mobilizing serum factors and immune cells through 831 exercise to counteract age-related changes in cancer risk. Exercise Immunology Review 26, 80-832 99. 833 834 Katzke VA, Kaaks R & Kuhn T. (2015). Lifestyle and cancer risk. Cancer Journal 21, 104-110. 835 836 Kim J-S, Taaffe DR, Galvão DA, Clay TD, Redfern AD, Hart NH, Gray ES, Ryan CJ, Kenfield SA, 837 Saad F & Newton RU. (2023). Acute effect of high-intensity interval aerobic exercise on serum 838 myokine levels and resulting tumour -suppressive effect in trained patients with advanced 839 prostate cancer. Prostate Cancer and Prostatic Diseases 26, 795-801. 840 841 Kim J-S, Taaffe DR, Galvão DA, Hart NH, Gray E, Ryan CJ, Kenfield SA, Saad F & Newton RU. 842 (2022). Exercise in advanced prostate cancer elevates myokine levels and suppresses in -vitro 843 cell growth. Prostate Cancer and Prostatic Diseases 25, 86-92. 844 845 Kim JS, Galvão DA & Newton RU. (2021). Exercise -induced myokines and their effect on prostate 846 cancer. Nature Reviews Urology 18, 519-542. 847 848 Kurgan N, Tsakiridis E, Kouvelioti R, Moore J, Klentrou P & Tsiani E. (2017). Inhibition of human 849 lung cancer cell proliferation and survival by post -exercise serum is associated with the 850 inhibition of akt, mtor, p70 s6k, and erk1/2. Cancers 9, 46. 851 852 Lauby-Secretan B, Scoccianti C, Loomis D, Grosse Y, Bianchini F & Straif K. (2016). Body fatness 853 and cancer--viewpoint of the iarc working group. The New England Journal of Medicine 375, 854 794-798. 855 856 Leung PS, Aronson WJ, Ngo TH, Golding LA & Barnard RJ. (2004). Exercise alters the IGF axis in 857 vivo and increases p53 protein in prostate tumor cells in vitro. J Appl Physiol 96, 450-454. 858 859 Lewandowska AM, Rudzki M, Rudzki S, Lewandowski T & Laskowska B. (2019). Environmental 860 risk factors for cancer - review paper. Annals of Agricultural and Environmental Medicine : 861 AAEM 26, 1-7. 862 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint 863 Li X, Li C, Zhang W, Wang Y, Qian P & Huang H. (2023). Inflammation And Aging: Signaling 864 Pathways And Intervention Therapies. Signal Transduction and Targeted Therapy 8, 239. 865 866 Liang CC, Park AY & Guan JL. (2007). In vitro scratch assay: A convenient and inexpensive method 867 for analysis of cell migration in vitro. Nature Protocols 2, 329-333. 868 869 Maxwell SE & Delaney HD. (2004). Designing experiments and analyzing data: A model comparison 870 perspective, 2nd ed. Lawrence Erlbaum Associates Publishers, Mahwah, NJ, US. 871 872 Mckendry J, Breen L, Shad BJ & Greig CA. (2018). Muscle morphology and performance in master 873 athletes: A systematic review and meta-analyses. Ageing Research reviews 45, 62-82. 874 875 Mctiernan A. (2008). Mechanisms linking physical activity with cancer. Nature Reviews Cancer 8, 876 205-211. 877 878 Mehanna LE, Boyd JD, Walker CG, Osborne AR, Grady ME & Berron BJ. (2025). Functional 879 assessment of migration and adhesion to quantify cancer cell aggression. Soft Matter 21, 2946-880 2957. 881 882 Michailidou Z, Gomez-Salazar M & Alexaki VI. (2022). Innate immune cells in the adipose tissue in 883 health and metabolic disease. Journal of Innate Immunity 14, 4-30. 884 885 Michelet X, Dyck L, Hogan A, Loftus RM, Duquette D, Wei K, Beyaz S, Tavakkoli A, Foley C, 886 Donnelly R, O’farrelly C, Raverdeau M, Vernon A, Pettee W, O’shea D, Nikolajczyk BS, Mills 887 KHG, Brenner MB, Finlay D & Lynch L. (2018). Metabolic reprogramming of natural killer 888 cells in obesity limits antitumor responses. Nature Immunology 19, 1330-1340. 889 890 Moore SC, Lee IM, Weiderpass E, Campbell PT, Sampson JN, Kitahara CM, Keadle SK, Arem H, 891 Berrington De Gonzalez A, Hartge P, Adami HO, Blair CK, Borch KB, Boyd E, Check DP, 892 Fournier A, Freedman ND, Gunter M, Johannson M, Khaw KT, Linet MS, Orsini N, Park Y, 893 Riboli E, Robien K, Schairer C, Sesso H, Spriggs M, Van Dusen R, Wolk A, Matthews CE & 894 Patel AV. (2016). Association of leisure-time physical activity with risk of 26 types of cancer 895 in 1.44 million adults. JAMA Internal Medicine 176, 816-825. 896 897 Moro-García MA, Fernández -García B, Echeverría A, Rodríguez -Alonso M, Suárez -García FM, 898 Solano-Jaurrieta JJ, López-Larrea C & Alonso-Arias R. (2014). Frequent participation in high 899 volume exercise throughout life is associated with a more differentiated adaptive immune 900 response. Brain, Behavior, and Immunity 39, 61-74. 901 902 Orange ST, Jordan AR, Odell A, Kavanagh O & Hicks KM. (2022). Acute aerobic exercise -903 conditioned serum reduces colon cancer cell proliferation in vitro through interleukin -6-904 induced regulation of DNA damage. International Journal of Cancer 151, 265-274. 905 906 Orange ST, Jordan AR & Saxton JM. (2020). The serological responses to acute exercise in humans 907 reduce cancer cell growth in vitro: A systematic review and meta -analysis. Physiological 908 Reports 8, e14635. 909 910 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Parekh NJ, Krouse TE, Reider IE, Hobbs RP & Ward BM. (2019). Type i interferon-dependent CCL4 911 is induced by a cgas/sting pathway that bypasses viral inhibition and protects infected tissue, 912 independent of viral burden. PLoS Pathogens 15, e1007778. 913 914 Piccirillo R. (2019). Exercise -induced myokines with therapeutic potential for muscle wasting. 915 Frontiers in Physiology 10, 287. 916 917 Rogers MA, Hagberg JM, Martin WH, 3rd, Ehsani AA & Holloszy JO. (1990). Decline in VO2max 918 with aging in master athletes and sedentary men. Journal of Applied Physiology 68, 2195-2199. 919 920 Rundqvist H, Augsten M, Strömberg A, Rullman E, Mijwel S, Kharaziha P, Panaretakis T, Gustafsson 921 T & Östman A. (2013). Effect of acute exercise on prostate cancer cell growth. PloS one 8, 1-922 9. 923 924 Severinsen MCK & Pedersen BK. (2020). Muscle–organ crosstalk: The emerging roles of myokines. 925 Endocrine Reviews 41, 594-609. 926 927 Shaw P, Dey Bhowmik A, Gopinatha Pillai MS, Robbins N, Dwivedi SKD & Rao G. (2025). Anoikis 928 resistance in cancer: Mechanisms, therapeutic strategies, potential targets, and models for 929 enhanced understanding. Cancer Letters 624, 217750. 930 931 Soares CM, Teixeira AM, Sarmento H, Silva FM, Rusenhack MC, Furmann M, Nobre PR, Fachada 932 MA, Urbano AM & Ferreira JP. (2021). Effect of exercise -conditioned human serum on the 933 viability of cancer cell cultures: A systematic review and meta-analysis. Exercise Immunology 934 Review 27, 24-41. 935 936 Sproston NR & Ashworth JJ. (2018). Role of c-reactive protein at sites of inflammation and infection. 937 Frontiers in Immunology 9, 754. 938 939 Thune I & Furberg AS. (2001). Physical activity and cancer risk: Dose -response and cancer, all sites 940 and site-specific. Medicine and Science in Sports and Exercise 33, S530-550; discussion S609-941 510. 942 943 Tilg H, Ianiro G, Gasbarrini A & Adolph TE. (2025). Adipokines: Masterminds of metabolic 944 inflammation. Nature Reviews Immunology 25, 250-265. 945 946 Urbano AM, Ferreira LM & Alpoim MC. (2012). Molecular and cellular mechanisms of hexavalent 947 chromium-induced lung cancer: An updated perspective. Current Drug Metabolism 13, 284-948 305. 949 950 Urbano AM, Rodrigues CF & Alpoim MC. (2008). Hexavalent chromium exposure, genomic 951 instability and lung cancer. Gene Therapy and Molecular Biology 12, 219-238. 952 953 Vainio H, Kaaks R & Bianchini F. (2002). Weight control and physical activity in cancer prevention: 954 International evaluation of the evidence. European journal of cancer prevention : the official 955 journal of the European Cancer Prevention Organisation 11 Suppl 2, S94-100. 956 957 Vichai V & Kirtikara K. (2006). Sulforhodamine B colorimetric assay for cytotoxicity screening. 958 Nature Protocols 1, 1112-1116. 959 960 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Welch DR & Hurst DR. (2019). Defining The Hallmarks Of Metastasis. Cancer Research 79, 3011-961 3027. 962 963 World Medical A. (2013). World Medical Association Declaration of Helsinki: ethical principles for 964 medical research involving human subjects. JAMA 310, 2191-2194. 965 966 967 968 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint Author Contributions: conceptualization, C.M.S., L.M.R.F., A.M.T. and A.M.U.; formal analysis, 969 C.M.S., J.P.M. and A.M.U.; investigation, C.M.S., J.P.M. and A.M.U.; resources, L.R., A.P., A.M.T. 970 and A.M.U.; writing —original draft preparation, A.M.U.; writing —review and editing, C.M.S., 971 L.M.R.F., A.M.T. and A.M.U.; visualization, C.M.S.; supervision, A.M.T. and A.M.U.; funding 972 acquisition, A.M.U. All authors have read and agreed to the published version of the manuscript. 973 Funding: This research was supported by grant 10/22 from Centro de Investigação em Meio 974 Ambiente, Genética e Oncobiologia (CIMAGO), Portugal, and institutional grants to QFM -UC from 975 Fundação para a Ciência e a Tecnologia (FCT), Portugal (UIDB/00070/2020, 976 https://doi.org/10.54499/UIDB/00070/2020 and UIDP/00070/2020, 977 https://doi.org/10.54499/UIDP/00070/2020). 978 Acknowledgments: The authors wish to thank all study participants, who generously volunteered their 979 time and efforts in participating in this study, and Dr Paulo J. Oliveira (Mitochondrial Toxicology and 980 Experimental Therapeutics Laboratory, Center for Neuroscience and Cell Biology, University of 981 Coimbra, Portugal), for the generous gift of the A549 cell line. 982 Institutional Review Board Statement: The study was conducted in accordance with the Declaration 983 of Helsinki and approved by the Ethics Committee of the Faculty of Sports Sciences and Physical 984 Education of the University of Coimbra (reference CE/FCDEF-UC/00062013). 985 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 23, 2026. ; https://doi.org/10.64898/2026.03.19.713009doi: bioRxiv preprint