Carl‭ , you mentioned that this is a body signal. What's the origin if the signal? Is it a biopotential signal such as ECG/EKG ?

Yes exactly, it is an ECG signal. Most of my circuit is not shown here, but I am also using a driven right leg circuit as Olin notes in his answer.

The 3rd reference electrode (the DRL electrode) changes your question a lot. If you have a reference electrode, then you don’t require AC-coupling of the ECG electrodes. What issue are you trying to address by high-pass filtering the ECG inputs (instead of, say, high-pass filtering the in-amp output)?

The DRL electrode also brings an electric safety question. How are you doing patient galvanic isolation?

Ag/AgCl electrodes used for ECG measurements have an intrinsic half cell potential of 223 mV, if the electrodes are completely new. If one electrode for some reason is more dry than the other, the difference between the half cell potentials for the two electrodes can maybe reach 100 mV. This causes a differential DC signal to exist across the in-amp's input terminal, severely limiting the possible gain without driving the in-amp to saturation. Galvanic isolation is obtained by powering the active components via batteries, and using right leg as circuit common. No contact with building ground whatsoever.

Carl‭ DC servo loop is one way to deal with the DC offset at the in-amp output. The thinking behind the DC servo loop is explained in this question. It should help increase the usable output range of the in-amp.

Carl‭ You mentioned in other comments that you're already aware of the more conventional ECG signal chains. What's your rationale for not wanting to use those kinds of topologies?