Lipoprotein(a) (Lp(a)) is increasingly recognized as an important causal and independent risk factor for cardiovascular diseases (CVD) including atherosclerotic CVD (ASCVD) and calcific aortic valve disease (CAVD). Roughly 20% of the global population has high levels of Lp(a) (defined as >30–50 mg/dL) which has been associated with increased CVD risk [1]. Lp(a) consists of a low-density lipoprotein (LDL)-like moiety attached to the unique apolipoprotein(a) (apo(a)). Apo(a) is composed of repeating kringle (K) domains (KIV1-KIV10; KV), followed by an inactive protease domain. KIV2 is present in a variable number of identically repeated copies. KIV10 contains a strong lysine binding site (LBS) important for binding of Lp(a) to biological substrates including fibrinogen/fibrin [2]. Interestingly, the strong LBS in KIV10 is also required for the covalent binding of pro-inflammatory oxidized phospholipids (OxPLs) to this K domain: mutations disrupting this strong LBS completely abolish OxPL addition [3,4]. It has been shown that Lp(a) is the preferential carrier of OxPL in plasma [5], and epidemiological data suggests that the OxPL modification of Lp(a) correlates closely with both Lp(a) levels and the risk attributable to this lipoprotein [6,7]. Mechanistic studies ex vivo and in vitro have directly implicated the OxPL associated with apo(a) in potentiating atherogenesis through stimulation of proinflammatory pathways that promote endothelial dysfunction [[8], [9], [10], [11]], monocyte recruitment and transmigration across the endothelial barrier [12,13], and macrophage apoptosis and cytokine release [4,14]. Although immunohistochemical studies have shown that apo(a) colocalizes with epitopes recognized by the anti-OxPL E06 antibody in human arterial lesions [15], the contribution of the OxPL on Lp(a) to atherogenesis has yet to be directly shown, and the molecular targets of Lp(a) and its associated OxPL in vivo, remain largely undefined.

The development of in vivo models to study Lp(a) is challenging. This reflects in large part the unique evolutionary history of the gene that encodes apo(a) (LPA), which is only present in Old World Monkeys, apes and humans [16]. Furthermore, only human apo(a) contains a functional strong LBS in apo(a) KIV10, and therefore OxPL coordination on apo(a) is not observed in other species [3]. For these reasons, transgenic (Tg) animal models expressing human Lp(a) are required. Several Tg apo(a) mouse strains have been characterized over the years [17]. However, the mouse models developed to date are problematic for various reasons: either the apo(a)/Lp(a) expression is well below pathogenic levels [[18], [19], [20], [21]], and/or the models lacked a transgene expressing human (apolipoproteinB-100) apoB-100 [18,22,23]. The latter is significant in that mouse apoB-100 is unable to covalently bind with human apo(a) to form bona fide Lp(a) particles [24].

The assessment of atherosclerosis in Tg apo(a)/Lp(a) mouse models to date has been, largely focused on lesion size and lysine-dependent co-localization of apo(a) with markers such as fibrin. While some studies reported more extensive aortic lesions in Tg apo(a) mice [18,22,23,25], others reported no effect of apo(a) on total lesion area [20,21]. For double-Tg mouse models also expressing human apoB-100, results have been challenging to interpret. Although no difference in lesion size in the double transgenic mice compared to singly transgenic (human APOB only) mice was reported [[19], [20], [21]], circulating Lp(a) levels were comparatively low (<50 mg/dL). In some cases, mRNA from the human APOB transgene underwent post-transcriptional editing in the liver [26] giving rise to particles containing apoB-48, which cannot covalently assemble with apo(a). One study was reported using a 17K apo(a) species and human apoB, with Lp(a) levels of approximately 50 mg/dL. Lp(a)-expressing mice had significantly increased aortic root plaque area compared to human apoB-100-only mice; however, the mice were made uremic by nephrectomy in this study [27].

Crucially, the above investigations did not seek to identify specific cell-types and plaque components impacted by Lp(a) and its OxPL modification or to assess whether a unique pathological signature could be detected in these mice. To overcome the limitations of previous Tg Lp(a) models, we have established a novel double-Tg mouse generated by crossing a mouse line expressing an apo(a) isoform containing 13 KIV repeats with mice expressing full-length human apoB-100. This model is unique in that the mice have high levels of plasma Lp(a), with the majority circulating in plasma as covalent Lp(a) particles. We characterized the complex lesions observed in these mice after 12 weeks of feeding with a high-fat, high-cholesterol (HFHC) diet in the context of hepatic Ldlr expression knockdown.