Introduction
Adipocytes secrete extracellular vesicles (EVs), but how they affect whole-body metabolism remains unexplored. This is due to technical limitations of tracking and isolating adipocyte-derived EVs in vivo¹. To address this, we have developed two Cre-dependent EV reporter mouse models: 1) The EGFP-EV reporter that labels EVs by CD9 fused to enhanced green fluorescent protein (EGFP), and 2) The NanoLuc-EV reporter labeling EVs by CD63 fused to Nanoluciferase (NanoLuc) and hemagglutinin (HA). We hypothesized that viral delivery of adiponectin-driven Cre to CD9-EGFP mice induces adipocyte-specific expression of EGFP. We also hypothesized that viral delivery of CD63-NanoLuc-HA to adiponectin-Cre mice induces adipocyte-specific expression of NanoLuc, creating a more sensitive reporter system.

Methods
To obtain our EGFP-EV reporter, we designed an adeno-associated viral (AAV) vector with adiponectin-driven Cre. CD9-EGFP mice (Mus musculus) were intraperitoneally injected with varying doses of AAV9 (5*10¹⁰, 5*10¹¹, and 1*10¹² viral genomes) to determine the optimal dose for efficient AAV delivery.
To obtain our NanoLuc-EV reporter, we designed an AAV vector harboring a CD63-NanoLuc-HA construct. As a control, we included a vector with an IgK signal peptide-NanoLuc-HA construct. Both vectors were packed in AAV9 capsid, and adiponectin-Cre mice (Mus musculus) were given a dose of 5*10¹¹ viral genomes by intraperitoneal injection.
After two weeks, mice were humanly sacrificed by terminal anesthesia (10 mg/kg xylazine, 50 mg/kg ketamine).
Organs such as adipose tissues and liver, and plasma were harvested and evaluated for adipocyte-specific expression of CD9-EGFP or NanoLuc and HA by western blotting and luciferase assays.

Results 
For the EGFP-EV reporter, CD9-EGFP was not detected in non-adipose tissues (n=20) or non-AAV-treated CD9-EGFP mice (n=4). CD9-EGFP expression in tissues was robust using a dose of least 5*10¹¹ viral genomes and was detected in brown adipose tissue (BAT) (n=16) and in white adipose tissue (WAT), including inguinal WAT (n=16), epididymal WAT (n=18) and mesenteric WAT (n=7). Unfortunately, CD9-EGFP was not detected in plasma (n=15). For the NanoLuc-EV reporter, a dose of 5*10¹¹ viral genomes was tested. The HA-tag was not detected by western blotting of tissue homogenates (n=1). NanoLuc activity was detected in plasma (n=1) and tissues (n=1), with highest levels in liver, mesenteric and epididymal WAT. For control AAV, the NanoLuc levels were higher across tissues compared to the NanoLuc-EV reporter, having the highest levels in epididymal WAT, BAT, and plasma (n=1).

Conclusion 
To obtain a proper adipocyte-specific EV reporter mouse model, the reporter system needs to be sensitive enough to track circulating EVs. Using NanoLuc as a reporter instead of EGFP seems crucial in this context. The control AAV, reflecting normal cellular release, serves well as a baseline for comparison. Intriguingly, the highest NanoLuc levels were observed in the liver of the NanoLuc-EV reporter. This raises questions about the interplay between adipose tissue and the liver, and whether hepatocytes take up adipocyte-derived EVs. In conclusion, our NanoLuc-EV reporter enables adipocyte-specific Cre-dependent EV labeling, providing a tool to study adipocyte-derived EVs in vivo to unveil their role in normal physiology and during metabolic disturbances.