We hypothesized that expression of immunity to gastrointestinal nematodes and its relationship with dietary metabolizable protein (MP) supply in ewes depended on body protein- but not on body fat reserves. Twin-bearing ewes were trickle infected with Teladorsagia circumcincta. Three mid-pregnancy feeding treatments were used, calculated to maintain body reserves (HH), maintain body protein- and lose body fat reserves (HL), or lose body reserves (LL), and followed by 2 iso-energetic, periparturient feeding treatments, calculated to provide either scarce (LP) or adequate (HP) amounts of MP. At the end of the mid-pregnancy feeding treatment, HL- and LL-ewes had a smaller backfat depth (∼body fat reserves) than HH-ewes, and LL-ewes had a smaller muscle depth (∼ body protein reserves) and higher faecal egg counts (FEC) than HH- and HL-ewes. Overall, LL-ewes had higher periparturient FEC than HH- and HL-ewes, and LP-ewes had higher periparturient FEC than HP-ewes. However, LL-ewes offered the LP feed had higher periparturient FEC than ewes on any of the other combinations of feeding treatments. Feeding treatments affected plasma pepsinogen but not plasma IgA. The results support the view that body protein- but not body fat reserves may overcome consequences of dietary MP scarcity on periparturient breakdown of immunity to gastrointestinal nematodes.