Lipins, Lipinopathies, and the Modulation of Cellular Lipid Storage and Signaling

2.4.1. Properties of lipin-1 isoforms generated by alternative splicing

The genes encoding human and mouse lipin-1 undergo alternative splicing to generate multiple protein isoforms known as lipin-1α, lipin-1β, and lipin-1γ [32, 41, 42]. The 890–amino acid sequence of the human lipin-1α isoform (891 amino acids in mouse) gains an additional in-frame exon of 36 amino acids (human) or 33 amino acids (mouse) to generate the lipin-1β isoforms, or a different in-frame exon of 26 amino acids to generate the human and mouse lipin-1γ isoforms. The inclusion of the lipin-1β or lipin-1γ–specific exons does not alter the PAP active site, the nuclear localization motif, coactivator motifs, or other known functional sites within the lipin-1 protein sequence. PAP activity studies on the three lipin-1 isoforms purified from an E. coli expression system indicated that the lipin-1α and lipin-1β isoforms have greater PAP specific activity by an order of magnitude than lipin-1γ, but that lipin-1γ has higher affinity for the PA substrate [42].

Intriguing observations have been reported regarding functional differences among lipin-1 isoforms, and the regulation of LPIN1/Lpin1 alternative splicing. Using an assay that allows the direct comparison of lipin-1α and lipin-1β mRNA levels across a panel of mouse tissues, it was determined that lipin-1α is more prominent in brain, spleen, and preadipocytes, whereas lipin-1 β is more prevalent in liver, heart, kidney, and mature adipocytes; levels of the two isoforms were similar in several other tissues (skeletal muscle, small intestine, lung, adrenal, ovary, and thymus) [32]. Thus far, the expression of the lipin-1γ isoform has been assessed in a limited number of tissues. It was shown to be expressed in human brain, but not liver, small intestine, or skeletal muscle [43]. There is some evidence that the three lipin-1 isoforms also exhibit differential subcellular localization. In adipocytes, lipin-1α tends to have predominantly nuclear localization, whereas lipin-1β is more often cytosolic [32]. The lipin-1γ isoform has been localized to lipid droplets, and expression of recombinant lipin-γ in COS-7 cells led to the appearance of small, clustered lipid droplets compared to larger, dispersed lipid droplets in the absence of this protein [43]. However, some caution is warranted in interpreting studies of lipin-1 isoform localization because antibodies to discriminate between different lipin-1 isoforms are not available and all work has been performed by protein overexpression in cultured cells.

Lipin-1α and -1β isoforms may have distinct functions in adipocyte lipid metabolism. When lipin-1–deficient adipocytes were complemented with these isoforms individually, lipin-1α promoted the expression of adipogenic genes (PPARγ, C/EBPα, aP2), whereas lipin-1β promoted the expression of lipogenic genes (fatty acid synthase, acetyl CoA carboxylase, stearoyl CoA desaturase, diacylglycerol acyltransferase) [32]. Evidence for distinct effects of lipin-1α and lipin-1β isoforms has also emerged from human studies. For example, increased expression of lipin-1β, but not lipin-1α, in adipose tissue was associated with protection from lipodystrophy in patients undergoing antiretroviral therapy for human immunodeficiency virus [44]. Lipin-1β, but not lipin-1α, was induced in adipose tissue of insulin-resistant subjects in response to treatment with the insulin-sensitizing drug pioglitazone, which also promotes increased adiposity [45]. Lipin-1β, but not lipin-1α, levels in the liver were increased in obese patients following gastric bypass surgery [46]. It is interesting to note that alternative lipin-1 splice variants are conserved in species beyond human and mouse, including pig and fruit fly. In pigs, the ratio of porcine lipin-1β to lipin-1α in skeletal muscle differs between animals with high and low intramuscular fat content [47]. In Drosophila, the single lipin-1 ortholog undergoes alternative splicing to produce at least three protein isoforms, which exhibit distinct tissue expression and subcellular localization patterns [48].

The regulation of lipin-1 mRNA splicing is also intriguing. During adipocyte differentiation, lipin-1β is induced preferentially over lipin-1α by glucocorticoids, suggesting that hormones regulate Lpin1 mRNA splicing [49]. Recently, the SFRS10 splicing factor has been implicated in the inclusion of the lipin-1β exon. SFRS10 was identified as showing differential expression levels in liver and muscle of lean and obese individuals [50]. SFRS10 was subsequently shown to regulate LPIN1 splicing in an in vitro minigene assay, and SFRS10 knockdown in HepG2 cells led to increased lipin-1β expression relative to lipin-1α [50]. The authors suggested that the down-regulation of SFRS10 in the liver of obese individuals may lead to increased lipin-1β expression and contribute to hepatic steatosis in the setting of obesity [50]. However, in a subsequent study, SFRS10 or lipin-1β expression levels in human liver biopsies did not appear to correlate with obesity or liver histopathology [51]. Technical and biological details differed between the two studies of SFRS10 levels [52], and more studies will be required to fully establish the roles of SFRS10 and lipin-1β in obesity.

2.4.2. Lipin-1 protein phosphorylation and sumoylation

The subcellular localization of lipin proteins is dynamically regulated and is likely a determinant of their activity. It has been known for many years that fatty acid availability promotes the migration of rat liver PAP activity from the cytosol to microsomal membrane fractions [27]. PAP enzyme subcellular localization appears to be regulated, at least in part, by posttranslational protein modifications of lipin proteins. Phosphorylation and sumoylation are the primary lipin-1 protein modifications that have been characterized to date.

Mouse lipin-1 is phosphorylated at multiple serine and threonine residues upon treatment of adipocytes with insulin or amino acids [31, 41, 53]. Insulin-stimulated lipin-1 phosphorylation is regulated, at least in part, by the mammalian target of rapamycin (mTOR) pathway and causes phosphorylation of Ser-106 on lipin-1 [53]. Lipin-1 and lipin-2 are both phosphorylated during mitosis at consensus sites for cyclin-dependent kinase 1 [54]. The yeast lipin ortholog, Pah1p, is phosphorylated by several kinases, including Cdc28p [55-57], and is dephosphorylated by a dual protein complex composed of Nem1p (a phosphatase) and Spo7p (a regulatory unit) [55]. The mammalian counterpart of this phosphatase complex has now been identified as CTDNEP1 (C-terminal domain nuclear envelope phosphatase 1, formerly Dullard) [58] and NEP1-R1 (nuclear envelope phosphatase 1-regulatory subunit 1) [59]. The CTDNEP1–NEP1-R1 complex can dephosphorylate lipin-1α, lipin-1β, and lipin-2 in cultured cell lines [59]. As a consequence of this dephosphorylation, lipin-1β localization to the nucleus increases. The tissue expression patterns of CTDNEP1 and NEP1-R1 in mice and humans mirror that of lipin-1 [59], suggesting a potential relationship between this complex and lipin-1 activity. Further studies are required to establish the role of the phosphatase complex in the function of lipin proteins (and potentially other proteins) in vivo. Lipin-1 is also modified by sumoylation in neuronal cells [30]. Sumoylation promotes the nuclear localization of lipin-1α, and intact sumoylation sites are required for lipin-1α coactivator activity [30].

Initial studies indicated that lipin-1 phosphorylation status primarily affects lipin-1 subcellular localization: highly phosphorylated lipin-1 is enriched in the cytosol and dephosphorylated lipin-1 is enriched on ER/nuclear membranes [31, 53]. Somewhat surprisingly, the phosphorylation status in these cases did not directly affect PAP activity as assessed in an in vitro assay performed with micelles containing the PA substrate. Very recent work by Harris, Brindley and colleagues has provided important insight into the relationship between lipin-1 phosphorylation and PAP enzyme activity. By manipulating the presentation of PA substrate in liposomes containing various amounts of phosphatidylcholine and phosphatidylethanolamine, these investigators demonstrated that lipin-1 enzymatic activity is influenced by the phospholipid composition of membranes containing PA, as well as by pH [60]. The proposed mechanism is that neighboring membrane phospholipids and cellular pH both influence the ionization status of PA, which in turn affects lipin-1 binding and PAP activity. The phosphorylation status of lipin-1 influences its ability to recognize di-anionic PA [60], and this interaction is mediated through the polybasic domain on lipin-1 (previously shown to influence both nuclear localization and PA binding [31, 33]. Thus, current knowledge paints a complex picture in which the regulation of lipin-1 PAP activity is influenced by multiple metabolic variables, including intracellular pH, the phospholipid composition of the ER membrane, and hormones or specific kinases and phosphatases that influence lipin-1 protein phosphorylation.

2.4.3. Lipin protein oligomer formation

Harris and colleagues demonstrated that lipin proteins form homo- and hetero-oligomeric structures [38]. Endogenous lipin-1 isolated from 3T3-L1 adipocytes—as well as recombinant lipin-1 purified from an insect cell expression system—self-associates as dimers and tetramers. Interestingly, they appear to associate in a head-to-head/tail-to-tail manner through N-LIP/N-LIP and C-LIP/C-LIP interactions on lipin-1 proteins. Lipin-1 can also form hetero-oligomers in combination with lipin-2 or lipin-3. Curiously, however, the formation of lipin-1 oligomers is not required for, nor does it enhance, lipin-1 PAP activity [38]. The purpose of lipin oligomer formation remains unknown. One suggestion is that oligomerization influences lipin protein subcellular localization. This is supported by the demonstration that lipin-1 protein that was targeted to the plasma membrane with an isoprenyl lipid anchor attracted wild-type lipin-1 molecules (i.e., proteins lacking lipid anchors) to the plasma membrane, presumably through lipin–lipin protein interactions [38]. It will be important to further investigate the physiological relevance of lipin homo- and hetero-oligomer formation.

3.2.1. Lipin-1 physiological functions revealed in mouse models

The fld mouse carries a complex gene rearrangement resulting in a null mutation and a variety of metabolic defects. A point mutation in the Lpin1 gene leading to a single amino acid substitution (Gly84Arg) was identified in the fld^2J mouse strain and confers the same phenotypes observed in fld mice [11]. The fld mutation was named “fatty liver dystrophy” based on the presence of a severe fatty liver that is evident during the neonatal period until about 2 weeks of age, and a progressive peripheral neuropathy that develops in young adult mice [9, 64]. The fld mice have lipodystrophy characterized by impaired adipocyte differentiation and a near-complete absence of TAGs in white adipose tissue depots [10, 11, 66-68]. Lipin-1–deficient mice have dramatically altered diurnal metabolism and are insulin resistant, likely related to dysregulated lipid metabolism in adipose tissue, liver, and muscle [69, 70]. Finally, male fld mice are infertile, despite producing motile spermatozoa [9].

The role for lipin-1 as the primary PAP activity in adipose tissue [26, 53] provides a partial explanation for the dramatically reduced TAG stores in adipose tissue of fld mice. The requirement for lipin-1 in adipose tissue ha s been further elucidated by studies of adipose tissue–specific lipin-1 knockout mice, which also manifest lipodystrophy, confirming the cell autonomous function of lipin-1 in white and brown adipocytes [67]. In addition, the adipose tissue–specific expression of a truncated lipin-1 that retains co-regulator function but lacks PAP activity confirmed the requirement for lipin-1 PAP activity in adipocyte maturation and TAG storage [66]. It is possible that the loss of lipin-1 coactivator activity and dysregulation of hepatic fatty acid oxidation genes [34] contributes to the neonatal fatty liver phenotype in fld mice. However, the recent demonstration that neonatal hepatic steatosis occurs in mice with lipin deficiency exclusively in adipose tissue [66] suggests that the fatty liver may result from metabolic dysfunction in adipocytes and/or macrophages.

The peripheral neuropathy in fld mice is characterized by demyelination and lipoatrophy in peripheral nerves [64, 71]. Inactivation of lipin-1 function specifically in Schwann cells revealed that it is the loss of lipin-1 function in that cell type, rather than in the adipocytes of the epineurium, that is responsible for the neuropathy [72]. Furthermore, Chrast and colleagues determined that the lack of lipin-1 PAP activity in Schwann cells leads to an accumulation of the PAP enzyme substrate, PA, and activation of MEK-Erk signaling pathway, which inhibits Schwann cell differentiation [72]. PA accumulation also occurs in other cell types of lipin-1–deficient mice, including heart [73] and adipose tissue [66, 67, 74]. In adipocytes, PA accumulation and Erk activation inhibits the expression of PPARγ, a key transcriptional regulator of adipogenesis, providing an explanation for the impaired adipocyte differentiation in fld mice [74]. The accumulation of PA in adipose tissue also causes impaired adipocyte lipolysis through allosteric effects on phosphodiesterase 4 activity [66].

The cause of male infertility in fld mice has not been investigated extensively, but this phenotype may be a behavioral change related to the peripheral neuropathy. However, recent findings implicate lipin-1 as a determinant of nuage structure in gonads [75], which could potentially affect testis function. Nuage is a structure that occurs between aggregated mitochondria and is implicated in the generation of small non-coding Piwi-interacting RNAs (piRNAs) that are expressed in animal gonads and are thought to influence retrotransposase-induced DNA damage, mitochondrial fission/fusion dynamics, and other processes [76]. Interestingly, the testes from fld mice exhibit abnormal nuage density and length [75].

3.2.2. The role of lipin-1 in muscle is highlighted in human lipin-1 deficiency

The first human lipin-1–deficient patients were identified in 2008 [13]. LPIN1 mutations were in children from consanguineous families who had experienced episodes of rhabdomyolysis with muscle pain, myocyte degradation, and fever. Homozygosity mapping directed attention to a region on chromosome 2, and sequencing revealed unique pathogenic LPIN1 mutations in multiple families [13]. Michot and colleagues have subsequently cataloged LPIN1 mutations in patients with childhood rhabdomyolysis from around the world [77, 78]. To date, at least 19 distinct LPIN1 mutations have been identified, most of which are point changes or small insertion/deletions that result in premature termination codons. In addition, an intragenic deletion of 73 amino acids has been reported in numerous unrelated patients [77, 78]. The occurrence of this mutation in multiple families suggests a founder effect within an isolated population.

LPIN1 mutations account for a significant proportion of childhood rhabdomyolysis in cases where fatty acid oxidation defects have been excluded. Michot and colleagues sequenced LPIN1 in 141 cases of severe or moderate childhood rhabdomyolosis [78]. Biallelic LPIN1 mutations were present in two-thirds of the individuals presenting with severe rhabdomyolysis with onset before 7 year s of age, but were infrequently observed in individuals with later onset or less severe rhabdomyolysis (Fig. 4). Interestingly, heterozygous LPIN2 and LPIN3 mutations were identified in a handful of patients with moderate rhabdomyolysis or with severe rhabdomyolysis with late onset (Fig. 4). Given that these mutations were observed only in the heterozygous state and that LPIN2 and LPIN3 are expressed at much lower levels in muscle than LPIN1 [26], it is unclear whether these mutations are truly related to the development of rhabdomyolysis.

Lipin-1 deficiency can produce dramatic, life-threatening illness in toddlers, with severe cardiac muscle damage leading to cardiac arrest [79]. In addition to the evidence for rhabdomyolysis, analysis of muscle lipids in two affected individuals revealed increased levels of lipid intermediates in the glycerol phosphate pathway [13]. An in-depth analysis of muscle metabolism in lipin-1–deficient mouse models could prove valuable in elucidating the basis for rhabdomyolysis in lipin-1–deficient individuals. In mouse skeletal muscle, lipin-1 provides virtually all of the PAP activity [26, 53], and muscle lipin-1 levels affect energy metabolism [80] and mitochondrial content [81]. Although rhabdomyolysis and muscle pathology have not been noted in earlier descriptions of fld mice, this may be related to difficulties recognizing episodic muscle pain in mice and a lack of stimuli that normally trigger those episodes in affected humans (e.g., infection/fever, strenuous exercise, fasting). Indeed, preliminary studies in our laboratory indicate that, under metabolic stress, lipin-1–deficient mice exhibit histolopathologic changes in muscle and elevated creatine kinase levels (Zhang and Reue, unpublished). Thus, lipin-1–deficient mice will likely be useful in better understanding muscle disease in lipin-1–deficient humans.

As described earlier, the importance of lipin-1 in adipose tissue in rodents is well established. It is therefore surprising that lipin-1–deficient children do not exhibit overt lipodystrophy [13]. Nevertheless, there is strong evidence that lipin-1 is a determinant of adiposity and related metabolic traits in humans. Both lipin-1 expression levels in human adipose tissue and LPIN1 genetic polymorphisms are associated with body mass index and insulin sensitivity in multiple human cohorts [17, 82-87]. Furthermore, most of the reports of human lipin-1 deficiency have been in children, and it remains possible that some affected patients will develop adipose tissue phenotypes as they age. It is also possible that metabolic differences between mice and humans may protect lipin-1–deficient humans from overt lipodystrophy. These include a greater reliance on de novo lipogenesis in mouse adipose tissue than in human adipose tissue [88], and/or potential compensation for the loss of lipin-1 in humans by other lipin family members. A better characterization of adiposity in lipin-1–deficient patients, and of the role of lipin-1 in the differentiation of human adipocytes, would be useful.

Whether lipin-1–deficient humans develop fatty liver during the first few weeks of life is unknown. However, it is interesting that a genetic variation in LPIN1 has been linked to the presence and the severity of fatty liver disease in children [89].

3.3.1. Lipin-2 and lipin-3 molecular function and regulation

Since the identification of the three lipin genes in mammals, lipins have been the subject of >180 publications, but fewer than 10% of those have dealt with lipin-2 or lipin-3. Both lipin-2 and lipin-3 are PAP enzymes, albeit with lower specific activities than lipin-1 [26]. Lipin-2 and lipin-3 carry the coactivator motif found in lipin-1 [34], and lipin-2 exhibits coactivator activity in a cell-based assay [61]. Lipin-2 is prominently expressed in liver [61], and knockdown studies have shown that lipin-2 contributes substantially to PAP activity and TAG synthesis in hepatocytes [62]. Lipin-2 expression levels in mouse liver are induced by fasting, diet-induced obesity, and ER stress [61, 62, 90], suggesting that modulating lipin-2 levels may be important for hepatic lipid homeostasis. Lipin-2 is also expressed in macrophages and protects against saturated fatty acid-induced inflammation in these cells in vitro [91]. Lipin-2 appears to differ from lipin-1 in its subcellular localization. While data on the localization of lipin-2 in tissues is not available, lipin-2 localizes primarily to the cytosol in HeLa cells, whereas lipin-1 is found in both the cytosol and an ER/nuclear membrane–associated pool [54].

Virtually no studies have addressed the physiological function of lipin-3. However, genome-wide association studies in humans have identified genetic variation in LPIN3 (as well as LPIN2) as determinants of metabolic traits. A variant near LPIN3 has been associated with fasting serum glucose levels in healthy individuals [92]. A polymorphism in the 3′ UTR of LPIN2 has been linked to fat distribution, glucose, and insulin levels in type 2 diabetics [93]. Furthermore, LPIN2 null mutations cause the rare disease, Majeed syndrome. Studies of lipin-2–deficient humans and mice have begun to define some of the physiological roles of lipin-2 (see sections 3.3.2. and 3.3.3).

3.3.2. Majeed syndrome

In 2005, Ferguson and colleagues identified the first mutations in LPIN2 in children from two families in Jordan [12]. The initial symptoms are usually fever and pain in bones and joints, typically beginning in the first year of life [94]. Radiography reveals osteomyelitic lesions near the ends of long bones and other locations, but these lesions are sterile [94-96]. In addition, subjects develop congenital dyserythropoietic anemia of varying severity [12, 96]. Some patients develop transient inflammatory dermatosis, but this is not a consistent disease phenotype.

Majeed syndrome is rare, having been observed in only four families, all from the Middle East [97-99]. Four LPIN2 mutations are known, three frameshift or splice site mutations that result in premature stop codons, and one missense mutation that changes a conserved serine to a leucine (Ser734Leu). The latter mutation occurs downstream of the PAP active site and coactivator motif and is conserved in mammalian lipin-1, -2, and -3 [12]. The Ser734Leu mutation abolishes PAP enzyme activity but does not impair the ability of lipin-2 to associate with microsomal membranes or adversely affect lipin-2’s transcriptional coactivator function [61]. Of note, lipins in all multicellular organisms have a serine at the corresponding position, whereas single-celled organisms have a threonine. Thus, a hydroxyl group at this position may be critical for their phosphatase activity. In any case, it is clear that the loss of lipin-2 catalytic activity is sufficient to cause Majeed syndrome.

Due to the rarity of Majeed syndrome and the limitations on performing studies in human subjects, mysteries remain regarding the pathogenesis of the disease and how it relates to lipin-2 function and glycerolipid metabolism. A recent report implicated interleukin-1β (IL-1β), an inflammatory cytokine, in the bone abnormalities in Majeed syndrome. Two Majeed patients showed clinical improvement when treated with drugs that block IL-1 action (Anakinra, an IL-1 receptor antagonist, and Canakinumab, an IL-1β antibody) [99]. Therefore, it seems possible that the bone abnormalities in Majeed syndrome are due to cytokine elaboration, either systemically or locally in bone. Cytokine production does occur within joints in disorders such as osteoarthritis [100, 101]. Lipin-2 mRNA expression is detectable in mouse bone and bone marrow [61], so further studies in mice could prove to be useful for defining the pathogenesis of Majeed syndrome (see section 3.3.3).

3.3.3. Lipin-2–deficient mice reveal cooperative and unique functions of lipin-1 and lipin-2

To investigate the functions of lipin-2, a Lipin-2–deficient mouse model was generated by targeted trapping [102]. Based on the prominent lipin-2 expression in liver [26, 62] and the substantial reduction in PAP activity in isolated hepatocytes with lipin-2 knockdown [62], it was expected that these mice would exhibit dramatically reduced liver PAP activity and impaired hepatic lipid metabolism. Unexpectedly, however, lipin-2 knockout (KO) mice have slightly elevated hepatic PAP activity and had a compensatory increase in lipin-1 protein levels [102]. As a result, hepatic and plasma lipid levels in chow-fed lipin-2 KO mice are nearly normal. However, when mice are challenged with a high-fat diet, Lipin-2–deficient mice accumulate excess hepatic TAG. Adenoviral knockdown of lipin-2 in wild-type mouse liver led to increased expression of ER stress genes [90], but no evidence for ER stress was detected in the liver of Lipin-2–deficient mice [102].

Observation of the Lipin-2–deficient mice as they aged revealed an unexpected role for lipin-2 in lipid homeostasis in the cerebellum. Beginning at ~6 months of age, Lipin-2–deficient mice develop a tremor and ataxic gait [102]. An analysis of lipin protein levels in the cerebellum of wild-type mice revealed that both lipin-1 and lipin-2 (but not lipin-3) are abundant in the cerebellum of young animals, suggesting that lipin-1 compensates for the lack of lipin-2 in the cerebellum of young Lipin-2–deficient mice. However, as mice age, lipin-1 levels in the cerebellum fall to nearly undetectable levels. Thus, after 6 months of age, the cerebellum of Lipin-2–deficient mice have virtually no lipin protein. This leads to an accumulation of PA. In the cerebellum, lipin-2 is normally localized to Purkinje cells [102]. Aging Lipin-2–deficient mice exhibit a substantial reduction in Purkinje cell number, implying that lipin-2 and glycerolipid synthesis may play a role in maintenance of that cell type. Cerebellar disease has not been reported in Majeed syndrome, but all of the reported patients have been under 25 years of age. It seems possible that Majeed patients might develop cerebellar disease later in life, and it is also possible that lipin-2 levels in the cerebellum could prove to be relevant to the pathogenesis of idiopathic late-onset cerebellar ataxias.

Like Majeed patients, Lipin-2–deficient mice develop dyserythropoietic anemia [102]. It will be important to determine if erythrocytes and hematopoietic tissues from these mice have PA accumulation and/or abnormalities in phospholipid content. It is difficult to ascertain whether lipin-2 KO mice experience episodic bone pain, as in Majeed syndrome. No bony lesions were detected in Lipin-2–deficient mice by routine radiography or micro-computed tomography of tibiae [102]; however, conventional radiographs can be normal in Majeed patients early in the disease course [94, 97]. Detecting bone lesions in lipin-2 KO mice may require more sensitive techniques (e.g., isotope uptake) [96] or studies of older mice. Monitoring mice for markers of inflammation or fever would also be informative.

The availability of both lipin-1–deficient and Lipin-2–deficient mouse strains made it possible to determine whether mice lacking both of these proteins were viable. It was determined that deficiency in both lipin-1 and lipin-2 is lethal during embryonic development, with affected embryos failing to develop a functional circulatory system [102]. This finding indicates that the presence of lipin-1 or lipin-2 is required during embryogenesis, and that lipin-3 cannot substitute for these functions.

The characterization of Lipin-2–deficient mice uncovered physiological functions of lipin-2 in the liver under dietary stress, in the cerebellum, in erythrocytes, and in embryonic development [102]. More generally, these studies have begun to illuminate physiological relationships between members of the lipin protein family, which operate in a tissue- and age-specific manner. In the case of lipin-1 and lipin-2, it is clear that each of these proteins has unique roles in specific tissues, while other physiological functions are shared by both lipin-1 or lipin-2 (Fig. 5). Further analysis of mice with global and tissue-specific deficiencies in the lipin proteins will likely enhance our understanding of the physiological relationships between lipin-1 and lipin-2.