**Background:** Feed restriction occurs frequently during pig growth due to economic or environmental stressors. Local breeds like Creole (CR) are suggested to have better tolerance to feed restriction than highly selected breeds like Large White (LW), but the molecular mechanisms are poorly understood. This study aimed to (1) compare the blood transcriptome response to feed restriction and refeeding in LW and CR pigs, and (2) investigate the effect of moderate feed restriction and refeeding on whole blood transcriptome.
**Methods:** Forty-four growing pigs (23 LW, 21 CR after 2 deaths) were housed in a semi-open building in Guadeloupe (tropical climate, mean 25.5°C ± 0.5°C, 88.4% ± 1.8% RH). The experiment had three periods: P1 (7 days, ad libitum for all), P2 (3 weeks, RF pigs had restricted feeder access 07:00–17:00, NF pigs ad libitum), and P3 (3 weeks, all ad libitum). Body weight, feed intake, rectal temperature (RT), and skin temperature (ST) were measured at the end of each period. Blood samples were collected at the end of P2 and P3. RNAseq was performed on blood from 28 animals (9 NF, 19 RF) using NovaSeq6000 S4, yielding ~48 million reads/sample. Reads were mapped to Sus scrofa 11.1 genome. DEGs were identified using DESeq2 (FDR < 0.05). Pathway analysis used ShinyGO (KEGG) and Ingenuity Pathway Analysis (IPA).
**Key Results:** During P2, RF pigs had lower ADFI (1.32 vs. 1.71 kg/day, P < 0.01) and ADG (475 vs. 766 g/day, P < 0.001) than NF pigs. Significant breed × treatment interactions were found: in LW, FE was lower in RF than NF (0.38 vs. 0.51, P < 0.01), while in CR, FE was similar between treatments (0.38 on average, P = 0.90). LW RF pigs had 12% lower BW after P2 (P < 0.001) and 10% lower after P3 (P < 0.01), while CR showed no significant BW differences between treatments. ST and RT were lower in RF pigs during P2 (34.6 vs. 35.7°C, P < 0.001; 39.1 vs. 39.4°C, P < 0.05). CR had lower ST than LW overall (35.6 vs. 36.4°C, P < 0.001).
Transcriptomically, after restriction (P2), CR had 648 DEGs (193 up, 455 down) vs. 198 in LW (62 up, 136 down), with 51 shared. After refeeding (P3), LW had 1538 DEGs vs. 187 in CR, with 47 shared. Within-treatment analysis (P3 vs. P2 in RF) showed 2087 DEGs in LW and 1375 in CR. KEGG enrichment after restriction was only significant for CR (39 pathways, mainly immune-related). IPA after restriction identified 179 canonical pathways for CR (top: Interferon Signaling, -log p=7.34, z-score=-3) and 29 for LW (top: Protein Ubiquitination Pathway, -log p=2.82). In LW, HSP genes (DNAJA1, DNAJC17, HSP90AA1, HSPA12B) were enriched in stress-related pathways. After refeeding, IPA found 30 pathways for LW (top: Oxidative Phosphorylation, -log p=3.44, z-score=1.414) and 27 for CR (top: Citrulline Biosynthesis, -log p=2.91). Intra-treatment RF analysis showed p38 MAPK activated in LW but not CR, and "cell survival" activated in both breeds in RF but not NF.
**Clinical Implications:** This study demonstrates that blood transcriptomics can dynamically capture breed-specific responses to nutritional stress. CR pigs, despite being more tolerant to feed restriction at the performance level, mounted a stronger transcriptional response during restriction, suggesting that transcriptional plasticity may underlie robustness. LW pigs showed greater transcriptional response to refeeding, with pathways related to thermogenesis and cell proliferation, consistent with their selection for high production. These findings highlight the value of local breeds as genetic resources for adaptation to climate change and suggest that breeding for robustness may involve selecting for enhanced transcriptional responsiveness to stress rather than minimizing it.