Hematopoietic stem cell transplantation-associated partial lipodystrophy

Abstract.

Hematopoietic stem cell transplantation (HSCT)-associated partial lipodystrophy (HSCT-PL) is a serious metabolic complication that develops in remote period among childhood cancer survivors treated with HSCT with total body irradiation (TBI). Since the first proposal in 2013, HSCT-PL seems to be increasingly recognized as a distinct disease entity. The patients with HSCT-PL show profound metabolic dysfunction including insulin resistance, diabetes, elevated triglycerides, and hepatic steatosis. Their body mass index is low–normal, although they show visceral fat accumulation and increased waist-to-hip ratio. In addition, HSCT-PL is characterized by Dunnigan phenotype: lipoatrophy in buttock and extremities combined with lipohypertrophy in face and neck. Although the precise pathogenesis is still obscure, radiation-induced damage to adipose progenitor cells, leading to accelerated senescence, seems to be a main pathway. Literature survey identified 17 patients of HSCT-PL with sufficient information from 12 reports. Among them, clear female predominance (15 females) and possible ethnic difference in disease prevalence (11 Japanese) were ascertained. Genetic factors may be involved in those epidemiological traits. There remains much to be clarified, including establishment of reliable diagnostic procedure, elucidation of long-term prognosis, and invention of effective treatment. Metreleptin is one of the promising options, and the accumulation of its therapeutic efficacy are warranted.

Highlights

● Hematopoietic stem cell transplantation (HSCT) causes metabolic late effects.

● HSCT-associated partial lipodystrophy is the severest end of metabolic sequelae.

● The concept of HSCT-associated partial lipodystrophy should be disseminated.

Introduction

Among childhood cancer survivors (CCS), those treated with hematopoietic stem cell transplantation (HSCT) are at particularly higher risk for developing late complications, compared with CCS without HSCT (1, 2). Accordingly, substantial numbers of CCS following HSCT suffer from a constellation of medical and psychological problems (3). Endocrinological late effects, such as growth failure, hypothyroidism, and hypogonadism, occur frequently, and have received much attention to promote early detection and pertinent intervention (3,4,5). Meanwhile, it has become clear that metabolic complication should be regarded as another significant late effect in CCS following HSCT, which include glucose intolerance, diabetes, insulin resistance, dyslipidemia, hepatic steatosis, metabolic syndrome, and sarcopenia (3, 6,7,8,9). HSCT-associated partial lipodystrophy (HSCT-PL) is one of these metabolic late effects found in CCS (10,11,12,13,14,15,16,17,18,19,20,21,22), the precise pathogenesis of which is still obscure.

More than a decade has passed since the author proposed the concept of HSCT-PL in Clinical Pediatric Endocrinology in 2013 (ref. 11, designated as “CPE paper” hereafter). Following the historical summary of the concept of HSCT-PL, this Review will summarize the clinical features of hitherto reported patients, try to elucidate the underlying pathology, as well as the discrepancies regarding gender and ethnicity, and to delineate the diagnostic challenges and the treatment strategies available at present and in future.

In this Review, HSCT-PL was arbitrary defined as the acquired, localized lipoatrophy eminent in buttock and extremities, which developed following any types of HSCT conducted during childhood or young-adult ages. In addition, presence of any metabolic dysfunctions, such as diabetes, dyslipidemia, hepatic steatosis, or insulin resistance was defined as required condition. The following factors were not considered to be indispensable: 1) Dunnigan phenotype, namely, lipoatrophy in buttock and extremities accompanied with lipohypertrophy in face and neck, 2) history of total body irradiation (TBI), 3) history of chronic graft versus host disease (cGVHD). No thresholds for metabolic laboratory test results and for the latent period between HSCT and HSCT-PL onset were set.

Process to the establishment of the disease concept

1) Proposal of disease concept

In 2013, the author reported five young adult patients following HSCT conducted to treat childhood malignancies (11). All of them showed peculiar subcutaneous fat distribution compatible with Dunnigan phenotype. They also shared profound metabolic dysfunctions such as severe insulin resistance, elevated triglycerides, and hepatic steatosis. Among them, patient 3, a male patient of 23 yr of age at that time, was a key patient who led to the concept linking HSCT and acquired partial lipodystrophy (APL) (Fig. 1). He had undergone two consecutive allogeneic HSCT for a relapsed infantile leukemia, the former one included TBI of 12Gy. He developed cGVHD mainly affecting joints and skin, that was treated with steroid and several kinds of immunosuppressants. After he entered puberty, he complained of thinning of extremities. Simultaneously, his Cushingoid appearance has become evident, which was suggestive of Dunnigan phenotype. At 17 yr of age, hepatic steatosis was diagnosed through the liver dysfunction and hepatic ultrasonography. At that time, high level of triglycerides was also noticed. Hyperinsulinemia was evident both in fasting state and following glucose loading, with HOMA-IR at 3.80 and maximal immunoreactive insulin (IRI) at 190 µIU/mL. Collectively, the diagnosis of partial lipodystrophy was made. Three other patients (patient 2, 4, 5) compatible with HSCT-PL could be identified by reviewing CCS patients that the author cared in Endocrinological unit. A while later, patient 1 was referred from Hematology unit, who presented with full-blown phenotype of HSCT-PL (11, 23). Then, a CPE paper was published describing 5 patients with HSCT-PL.

Fig. 1.

Patient 3 and Patient 4, previously presented in Ref (11). Patient 3 (Upper panel): Note the phenotypic change with the accumulation of fat tissue in cheek, giving rise the impression of Cushing-like appearance in advancing years. In addition, depicted here the thinness of extremities at his 23 yr of age. Patient 4 (Lower panel): Over 13 yrs no significant deterioration in carbohydrate and lipid metabolism was observed. BMI, body mass index; TG, triglycerides. Written informed consent for publication of facial photographs was obtained from both patients.

At that point, only one patient compatible with HSCT-PL was ascertained in literature: Rooney and Ryan reported a 23-yr-old female who developed partial lipodystrophy following HSCT for leukemia, presenting distinct Dunnigan phenotype (10). The patient suffered from sclerodermatous cGVHD. She later developed elevated triglycerides and diabetes necessitating insulin as much as 300 IU a day. Thus, the patient 3 in CPE paper and this patient shared many similarities except for the presence of overt diabetes.

2) Reports in medical literature

Hitherto reported patients are retrieved using PubMed and summarized in Tables 1 and 2 (10–22). Literature search was conducted in July 2025, using the keywords of “lipodystrophy”, “hematopoietic stem cell transplantation”, “total body irradiation”, and “childhood cancer survivors”. The reports without sufficient clinical information on each patient were excluded. Then, only the patients who met the above-mentioned definition of HSCT-PL were collected. Finally, the patients with known etiology for APL other than HSCT were excluded. As a result, from 12 reports, a total of 17 patients were collected. All patients are female, except for two male patients, namely, patient 3 and 4 in CPE paper. Another noticeable issue is the imbalance by country: Japanese patients constitute 11 patients, followed by three Italian patients. Most of the patients had leukemia as the underlying disease for HSCT, while two patients with neuroblastoma (patient 4 and 5 in CPE paper) and one patient with Hodgkin lymphoma. One patient had never undergone TBI (18), whereas a history of TBI was clearly stated in 14 patients. While cGVHD was ascertained in 13 patients, absence of cGVHD were documented in four patients, including patient 4 in CPE paper. Notably, among 13 patients with cGVHD, 8 patients had cGVHD involving skin. In accordance with previous reports, the interval between HSCT and the onset of HSCT-PL is almost a decade. Body mass index (BMI) of the patients distributed from low to normal range. The presence of Dunnigan phenotype, diabetes, and elevated triglycerides could be ascertained in 82%, 71%, and 94% of the patients, respectively. Severely insulin-resistant diabetes, defined as daily insulin requirements > 200 IU or > 2 IU/kg (24), was present in 5 patients, whereas severe hypertriglyceridemia > 1,000 mg/dL (25) was ascertained in 6 patients. Whereas leptin levels were distributed widely, adiponectin levels tended to be low.

Table 1. Reported patients of hematopoietic stem cell transplantation (HSCT) -associated partial lipodystrophy (PL) [1^st of two parts].

Table 2. Reported patients of hematopoietic stem cell transplantation (HSCT) -associated partial lipodystrophy (PL) [2^nd of two parts].

Two patients were excluded from Tables 1 and 2, because they did not meet the definition of HSCT-PL mentioned above. The first one was a woman of ‘generalized’ lipodystrophy’ following HSCT for relapsed T-cell lymphoma (26). The second patient developed APL as a rare manifestation of autoimmune polyglandular syndrome 1, more than 20 yr after HSCT conducted to treat multiple antibiotic-refractory fungal infections (27).

3) Current status

There was no reference to HSCT-PL in a landmark 2016 multi-society Practice Guideline (28). In 2021 guideline for lipodystrophy developed by Japan Endocrine Society, HSCT-PL was briefly mentioned in a section of “B-5: other types of acquired lipodystrophy”, alongside diencephalic syndrome and partial lipodystrophy of the limbs (29). Similar situation is observed in other guidelines and reviews from overseas. They only mentioned the reported cases of HSCT-PL, but seem to be reluctant to include HSCT-PL into APL as a definite entity (30, 31). Occasionally, HSCT-PL is not mentioned (32, 33). However, in a 2022 review of childhood lipodystrophy, “bone marrow transplant-associated lipodystrophy” is treated as a definite entity within APL, along with Barraquer-Simons syndrome, and HIV-associated lipodystrophy (34). The author is aware of two other reviews treating HSCT-PL as a single disease entity (35, 36). Furthermore, in a very recent European survey (ECLip registry cohort), a category of “APL associated with total body irradiation and hematopoietic stem cell transplant” was included in a section of APL, and the presence of seven cases were reported (37).

Suggested pathogenesis

Although the precise mechanism that will lead to the development of HSCT-PL is not yet fully elucidated, several hypotheses have been proposed. Fig. 2 shows the most plausible pathophysiology for HSCT-PL at present, which is a revision of a previously published one (23).

Fig. 2.

Suggested pathophysiology for hematopoietic stem cell transplantation-associated partial lipodystrophy (HSCT-PL). As one of the main pathogeneses for HSCT-PL, the hypothesis of total body irradiation (TBI)-induced damage in adipose progenitor cell (APC), with accelerated senescence, is most plausible. Reduced lipid-laden capacity in subcutaneous adipose tissue will lead to ectopic fat accumulation, while unhealthy expansion will occur in visceral fat. Chronic graft versus host disease (cGVHD), especially sclerodermatous cGVHD, may also damage subcutaneous adipose tissue. In this Figure, other possible pathogeneses are listed. In addition, clinical features in HSCT-PL obtained through the accumulated case reports are depicted.

1) Inferences by Rooney and Ryan claiming the role of cGVHD

As mentioned above, the first reported patient of HSCT-PL suffered from cGVHD that affected mainly skin, more precisely, epidermis, dermis, and dermo-epidermal junction (10). As the authors described the condition as sclerodermatous cGVHD, the patient showed grossly emaciated limbs. And the authors speculated that partial lipodystrophy was “consequent to the severe cutaneous injury she suffered as a result of sclerodermatous cGVHD.” In addition, they inferred that Dunnigan phenotype developed because lower legs, thighs, buttocks and forearms are the commonly affected areas during active cGVHD. This hypothesis seems to be valuable, considering that, out of 17 patients with HSCT-PL, 13 had cGVHD and 8 had dermal cGVHD. In addition, cGVHD involving skin will make the lipoatrophy more conspicuous, which may lead to the earlier recognition of abnormal fat distribution. Shibata et al. (14) proposed that HSCT-PL is one of the manifestations of cGVHD after HSCT. Kimura et al. speculated the fluctuation in cGVHD severity can directly affect the clinical course of HSCT-PL (12).

2) Limited expandability hypothesis

This hypothesis was initially presented by the author, emphasizing the damaging effect of TBI towards both subcutaneous and visceral adipocytes (11). According to this theory, damaged adipocytes will lose their capacity for lipid storage. By analogy with the spill-over theory, alternatively referred to Danforth’s hypothesis for type 2 diabetes (T2D), individuals with limited capacity of lipogenesis in adipocytes are incapable for storing lipids safely, and then the excess lipid will spill-over in ectopic organs such as liver, muscle, pancreas, and heart, which will give rise to insulin resistance (38,39). Indeed, adipose tissue obtained from mice after 7–10 Gy irradiation showed decreased numbers of mature adipocytes and proliferating cells, as well as significantly increased apoptotic cell counts (40). In addition, ob/ob mice that received 8 Gy TBI developed reduced body-fat mass with more severe insulin resistance and hepatic steatosis compared with controls (41). In a recent study, irradiated adipocytes obtained from mouse showed elevated oxidative stress, unrestrained lipolytic activity, upregulation of senescence and inflammation pathways, and immune cell infiltration, all of which persists chronically (42). With clinical aspect, among the survivors of childhood leukemia, those with HSCT+TBI had substantially worse metabolic outcomes compared to those treated solely by chemotherapy (43). This theory is straightforward and based on both basic and clinical evidence. In addition, this is attractive because it can elucidate the process developing lipoatrophy in buttock and extremities as well as the profound metabolic dysfunctions in HSCT-PL.

3) Donor-cell migration hypothesis

Tews et al. added another piece for the mechanisms to develop HSCT-PL (44). Circulating adipocyte progenitors have been shown to infiltrate preferentially into visceral adipose tissue than subcutaneous adipose tissue (45). Therefore, the authors speculated that donor-derived mesenchymal stem cells might migrate more often into the recipient’s visceral adipose tissue, accumulate over time, and replace host adipocytes. Then, the donor-derived adipocytes may contribute to visceral-dominant fat distribution and may cause metabolic dysfunctions. Although the direct evidence is lacking, this theory is capable to elucidate the increased visceral adipose tissue and the long latent period for developing HSCT-PL

4) Lipodystrophic and sarcopenic phenotype

Wei et al. studied body composition of CCS following HSCT and found reduced subcutaneous fat volume, increased visceral and total fat mass, and reduced lean mass (43). They described this condition as overlapping features of lipodystrophy and sarcopenic obesity. In addition, they emphasized the significance of measuring waist-to-hip ratio as a screening tool for identifying metabolic dysfunction in CCS following HSCT. Also, they recommended the intervention by increasing muscle mass, as well as reducing fat deposition in muscle, to reduce their cardiovascular risks. They also conducted a systemic review and found that muscle mass, and probably muscle function, was significantly lower in HSCT+TBI group compared to healthy controls (46). Although their observation is not directly relevant to the pathogenesis for HSCT-PL, the contribution of sarcopenia to the development of metabolic dysfunctions should also be considered.

Compared to Dunnigan phenotype seen in familial partial lipodystrophy (FPLD) type 2 (FPLD2, OMIM # 151660) due to LMNA gene mutation, thinness of extremity in some patients with HSCT-PL, as in patient 3 in CPE paper (Fig. 1), seems to be more striking (10, 14, 20). This difference may be caused by the reduction not only in adipose tissue mass but also in muscle volume: as stated earlier, presence of sclerodermatous cGVHD may even exaggerate the thinning of extremities.

5) Adipose progenitor damage hypothesis

Mayson et al., in their report of two patients with severe insulin resistance and hypertriglyceridemia after HSCT with TBI, speculated that preadipocyte may be damaged by radiation (6). In addition, some authors also referred to the radiation-induced damage on preadipocytes or adipose stem cells (13, 44, 47).

Today, adipocyte progenitors are collectively labeled as adipose progenitor cells (APCs), and they are subdivided into adipocyte stem cells, preadipocytes, and anti-adipogenic progenitors (48). Whereas adipocyte stem cells are specialized for cell proliferation, preadipocytes have a high adipogenic capacity. Anti-adipogenic progenitors, alternatively referred to fibro-inflammatory progenitors (49), suppress adipogenic differentiation. Fig. 3 depicts the lifelong variation in numbers of total-body mature adipocyte: from 14^th wk of gestation until just before puberty, except for initial 6 mo after birth, APC continues to proliferate steadily (48, 50,51,52), so the adipocytes’ number increase constantly. Following the rapid increase during adolescence, the adipocytes’ number stabilizes in adulthood (51). Adipocytes in adulthood are replaced at 10% rate annually (52). After middle age, around 40 yr of age, adipocytes’ number begins to decrease owing to reduced regenerative capacity of APC (48, 53). Therefore, adipose tissue is one of the first sites where senescence occurs (53).

Fig. 3.

Age-related changes in the numbers of mature adipocytes. Adipose progenitor cell (APC) starts to proliferate around 14^th gestational week. Natural course (black line): After birth, following initial 6 mo characterized by hypertrophy, APC continue to proliferate with peak rate at puberty. After 40 yr of age, the number of mature adipocytes begins to decrease owing to senescence of APC. In hematopoietic stem cell transplantation (HSCT)-associated partial lipodystrophy (red arrow): HSCT during childhood will impair the proliferative potency of APC, which leads to the decrease in numbers of mature adipocytes. The earlier the insult occurs, the more severe will be the total deficiency of adipocyte number.

HSCT procedure, TBI in particular, will damage the proliferative potency of APC, especially that of adipocyte stem cells. This will limit the number of mature adipocytes, which must become more prominent as time goes by (Fig. 3). Under such circumstance, decreased number in subcutaneous adipocyte will lead to insufficient capacity of lipid storage in skin, and then spill-over mechanism (Danforth’s hypothesis) is exerted (38,39). At the same time, in visceral adipose tissue, adipocyte size must increase because of limited number of mature adipocytes, which in turn leads to “unhealthy expansion” of fat depot. While healthy expansion denotes the combination of adipocytes with a variety of size, namely, mixture of hyperplasia and hypertrophy, unhealthy depot consists of large hypertrophic adipocytes. Such depot will be susceptible to mechanical stress, tissue hypoxia, inflammatory cell migration, fibrosis, and extracellular matrix accumulation, leading to adipose tissue dysfunction and eventually to comorbidities such as insulin resistance and T2D (48).

To support this hypothesis, the size of adipocytes obtained from gonadal white adipose tissue of irradiated mice was found to be increased (54). Heavy impact of irradiation on the proliferation and differentiation has been demonstrated in human primary preadipocyte (55) and in mesenchymal stromal stem cells obtained from normal adults (56). In addition, Visentin et al. analyzed the subcutaneous adipose tissue obtained from the patients with metabolic syndrome following HSCT and found the impairment in preadipocyte differentiation (57).

From this hypothesis, it may be suggested that HSCT-PL is a state of accelerated senescence process in APC. As stated above, irradiated murine adipocytes showed persistent upregulation of senescence pathways (42). This assumption is in accordance with the recent concept that CCS and adult cancer survivors are associated with accelerated aging, cellular senescence, and immune-senescence, the conditions resembling to those found in elderly (58,59,60).

6) Estrogen and growth hormone deficiency

Rajendran et al., in a report of a patient resembling HSCT-PL with marked hypertriglyceridemia, has suggested a role of estrogen deficiency for an impaired expanding capacity of subcutaneous adipose tissue (47). Although the patient had premature ovarian failure from 12 yr of age, it was stated that she was on Premarin^® treatment. Growth hormone (GH) status may also be relevant, because adults with GH deficiency are at risk of insulin resistance and increased visceral fat mass (61).

Gender difference

As listed in Table 1, HSCT-PL shows robust female predominance. Male patients with detailed information are limited to patient 3 and 4 in CPE paper. This gender difference is not surprising, because higher female prevalence, or more prominent symptoms in women, is observed in other disorders affecting adipose tissue, such as FPLD2, APL (Barraquer-Simons syndrome), and lipoedema (alternatively called as lipedema) (28, 29, 34, 62, 63). Even in simple obesity, clear gender differences exist in fat distribution: women tend to present with apple-shaped phenotype whereas men with pear-shaped (64). Size of adipocyte and adipose tissue functions are also influenced by gender (64). Considering that menopaused women show fat distribution similar to men, estrogen must play a pivotal role in gender differences stated above (64). In caveolae in mature adipocytes, estrogen receptor-α (ERα) is expressed, and ERα communicate with other proteins such as caveolin-1 and matrix metalloproteinase-14 (MMP-14) (63). Detailed mechanism how ERα modulates the size and function of adipocytes seems to be an open question at present (65).

Ethnic difference

Six out of 12 reports were from Japan, and 11 out of 17 patients were Japanese (Table 1). The second most frequently reported country was Italy. Ethnic difference has not been reported in FPLD2 and APL (Barraquer-Simons syndrome). The predominance of Japan and Italy in HSCT-PL may come from higher awareness of disease, because earlier reports, including CPE paper, were from Japan and Italy (13). Other possible factors include a difference in treatment for malignancies, patient follow-up policy, and publication bias. In addition, genetic factor(s) may have a role, considering that lipoedema, which is characterized by localized fat hypertrophy and may be regarded as a mirror image of partial lipodystrophy, often develop in Caucasian women. Although the precise mechanism of developing lipoedema has not been clarified, contribution of genetic factor, such as a polymorphism in caveolin-1, is under investigation (63).

In accordance with this assumption, recent studies suggested that Asian population may be more prone to lipodystrophic phenotype. 2020 study conducted by Indian researchers concluded that “thin fat phenotype” in south Asian people, indicating that they tend to develop T2D by smaller increase in BMI compared to white people, resembles partial lipodystrophy (66). The authors evaluated gene expression in the adipose tissue obtained from Indian patients with T2D, and found that many genes related to lipodystrophy, including LIPE (encoding hormone sensitive lipase, responsible for FPLD type 6, OMIM # 615980), CAVIN1 (encoding caveolae associated protein 1, responsible for congenital generalized lipodystrophy type 4, OMIM # 613327), PLIN1 (encoding perilipin 1, responsible for FPLD type 4, OMIM # 613877), and LMNA, showed significantly increased expression levels compared with non-diabetic counterparts (66). 2025 study utilizing partitioned polygenic score compared Pakistani/Bangladeshi with Europeans, and found that genes related with lipodystrophy, including PNPLA3 (encoding adiponutrin) and PPARG (encoding peroxisome proliferator-activated receptor gamma, responsible for FPLD type 3, OMIM # 604367), were the second most strongly associated with the development of T2D in south Asians after insulin-related genes (67).

Thus, there seem to be a possibility that the development of HSCT-PL is influenced by ethnicity. However, this will be proven only through the comparison of the epidemiological studies among diverse countries/regions.

Diagnostic challenge

Knowing the concept of HSCT-PL, it may not be necessarily difficult to suspect it during follow-up of CCS. In particular, those who had undergone HSCT during early childhood, those with complicated clinical courses such as metastases, relapsing episodes, or multiple HSCT procedures (68), and those with cGVHD, especially sclerodermatous cGVHD, may be regarded as high-risk group and should be under the enhanced clinical vigilance. In the patients with sclerodermatous cGVHD and/or profound sarcopenia, thinning of the extremities may be the first sign of HSCT-PL. Otherwise, the initial sign will be elevated triglycerides or abnormal liver function tests and/or imaging studies indicating hepatic steatosis. Highly insulin-resistant diabetes may also trigger the diagnosis of HSCT-PL, but the frequency of developing insulin-resistant diabetes is less frequent than that of elevated triglycerides and liver steatosis.

The obstacle for the definite diagnosis of HSCT-PL is the ascertainment of the abnormal distribution pattern of adipose tissue, owing to its subjective nature. Low adiponectin level will be helpful, but not sufficient. To address this issue, some objective indicators are highly needed. Imaging study, both CT and MRI, is useful to substantiate loss of subcutaneous adipose tissue, especially at buttock or lower legs. CT/MRI also helps to ascertain the increased visceral adipose tissue. Adiyaman et al. suggested that a pubic/gluteal fat ratio obtained by pelvic MRI is a promising method to diagnose FPLD with 96.7% sensitivity and 100.0% specificity (69). The utility of this index should be verified in the patients with HSCT-PL.

Differential diagnosis includes genetic counterpart that leads to Dunnigan phenotype, namely, FPLD (in particular, FPLD2 and FPLD3). In addition to detailed family history, careful examination for other family members may be necessary to rule out FPLD. Negative genetic testing for the candidate genes including LMNA and PPARG will be regarded as reliable basis for diagnosing HSCT-PL (70). In this situation, it should be kept in mind that circulating leukocytes are donor derived. Accordingly, DNA should be collected from recipient-derived tissues, such as fibroblast or nail tissues.

The most important issue to facilitate early diagnosis is to disseminate the concept of HSCT-PL among medical professionals who care for CCS.

Treatment

Therapeutic aspect of HSCT-PL has not been thoroughly studied. Accordingly, at present, treatment strategy should be personalized. Below, the possible treatment options, which are inferred from those in FPLD (28,29,30,31, 34, 35, 71) and from the reported cases with HSCT-PL, are described.

1) Lifestyle intervention

Lifestyle intervention will be a cornerstone for every patient with HSCT-PL, that includes medical nutrition therapy (72) and physical activity promotion (73). Patient 4 in CPE paper seems to underscore the importance of physical activity to prevent the rapid progression of metabolic dysfunctions in HSCT-PL (Fig. 1). At 20 yr of age, he showed diabetic pattern in OGTT with a peak IRI as high as 1,000 μIU/mL. However, he is free from medication for diabetes during the following 13 yr, with his recent HbA1c level at 6.1%. He walks more than 10,000 steps every day as a warehouse worker, and maintained his BMI around 18 kg/m². It seems that regular walking habit prevented the progression of diabetes, as well as hypertriglyceridemia, in this patient.

Content of diet should comply with the recommendation for T2D, although firm evidence is lacking. To set an ideal target weight may be sometimes challenging, because the patients with HSCT-PL have often short stature, reduced muscle mass, and, of course, abnormal fat distribution. Accordingly, the strategy to control body weight should be advised following careful evaluation in individual basis. Whether specialized form of medical nutrition therapy, such as isocaloric ketogenic diet or time-restricted feeding (72), is beneficial for the selected patients should be evaluated in future study.

2) Insulin sensitizers

Selection of antidiabetic agents other than insulin must be personalized. In patients with FPLD presenting diabetes, metformin is widely used as the first-line agent. From the viewpoint that HSCT-PL seems to be an accelerated senescence process in APC, it is noteworthy that metformin is considered to exhibit gerotherapeutic effects on multiple age-related diseases through the inhibition of senescence-associated secretory phenotype (74). Thiazolidinediones (TZD) may have a role (75), because TZD will exert insulin sensitizing effect and it can increase fat mass. Studies using TZD in lipodystrophy was summarized in a Practice Guideline (28). In addition, TZD was prescribed for some patients with HSCT-PL with variable effects (10, 14, 16, 17, 22). However, long-term treatment course with TZD has not been reported. Considering the cardiovascular concerns, it may be hesitated to make the patients with HSCT-LP under the continuous TZD usage.

3) Sodium-glucose cotransporter 2 inhibitors

Sodium-glucose cotransporter 2 (SGLT2) inhibitors have been demonstrated to reduce visceral adipose tissue mass (76). In addition, the degree of muscle mass reduction with SGLT2 inhibitors is believed to be minor compared to other antidiabetic agents (76). Because SGLT2 inhibitors are beneficial for hepatic steatosis, exert anti-inflammatory effects, and protect from cardiovascular complications, they may be one of the first-line therapeutic agents. Moreover, SGLT2 inhibitors reduced markers of senescent cells in adipose tissue of a mouse model of diabetes (74). The favorable effects of SGLT2 inhibitor, in combination with metreleptin, was reported in an “incomplete” HSCT-PL patient (16, 17). Of interest, in the reported patient, SGLT2 inhibitor exerted not only metabolic benefits, but redistribution of adipose tissue, namely, increasing visceral and subcutaneous adipose tissue mass (17).

4) Other anti-diabetic agents

Recently, the effectiveness of glucagon-like-peptide-1 receptor agonist (GLP1RA) in FPLD was reviewed (77). Experience of GLP1RA has also been reported in two patients with HSCT-PL (Table 2), with modest effects in controlling diabetes (19, 22).

In overt diabetes, exogenous insulin, sometimes with high doses, is necessary to overcome severe insulin resistance, as demonstrated in the first patient of HSCT-PL (10). In a patient with congenital generalized lipodystrophy (CGPL) accompanied with extreme insulin resistance, the long-term course with IGF-1 treatment was reported (78).

5) Metreleptin

The author has described the effectiveness of metreleptin in a patient 1 of CPE paper (23). During 28-mo of treatment with metreleptin, an absolute reduction of 4.2% in HbA1c level was achieved, accompanied with decreased HOMA-IR from 36.5 to 22.0. In addition, favorable effects were ascertained in liver function tests, triglycerides levels, and non-HDL cholesterol levels. Four additional reports are available, some of which suggested the usefulness of metreleptin in ameliorating metabolic dysfunctions with HSCT-PL (Table 2) (14,15,16,17, 21). Of interest, Nagayama et al. reported immediate anorexic effect of metreleptin (16), whereas patient 1 in CPE paper did not show any shift in her food intake (23). As seen in Table 2, the circulating leptin levels in the metreleptin-treated patients are widely distributed from 3.5–21.3 ng/mL (reference range for women: 2.5–21.8 ng/mL), which seems not relevant to the therapeutic effectiveness. The success in patient 1, with relatively higher leptin level (18.7 ng/mL), may indicate that metreleptin will work via pharmacological effect, not simply as a replacement.

Unsolved questions

1) Disease prevalence

Cross-sectional survey conducted in a Japanese local children’s hospital estimated the prevalence of HSCT-PL as high as 9.1% in CCS following HSCT, that is the only epidemiological study conducted ever (68). Because children’s hospital preferentially treats malignancy with severe clinical conditions, this prevalence may be an overestimation. In ECLip registry cohort stated above showed that among 64 patients with APL, seven patients were associated with HSCT-PL, whereas 40 were diagnosed as Barraquer-Simons syndrome and 17 had no known causes (37). As discussed above, prevalence of HSCT-PL may be diverse among various ethnicities. Therefore, region-specific prospective study is essential to estimate the real prevalence. In addition, establishment of common diagnostic criteria including some subjective indices must be of significance. In this sense, an ongoing national survey of HSCT-PL among Japanese population, led by the study group in Japan Endocrine Society, must have much significance.

2) Why Dunnigan phenotype?

The peculiar combination of lipoatrophy and lipohypertrophy in HSCT-PL is impressive. This pattern has much similarity with Dunnigan phenotype in FPLD2, a genetic disorder caused by mutations in LMNA gene encoding lamin A/C. Interestingly, another type of APL (Barraquer-Simons syndrome), that often accompanies nephropathies and low circulating C3 level, leads to contrasting fat distribution: lipoatrophy in face and lipohypertrophy in lower limbs (79). At present, the pathology behind Dunnigan phenotype in FPLD2 remains undetermined. Mutations in LMNA gene lead to variety of disorders other than FPLD2, such as cardiomyopathy, muscular dystrophy, progeria, and so on, collectively called laminopathy (80, 81). Each disorder has specific pattern of mutation spectrum: mutations in the tail or Ig-like domain are associated with lipodystrophy, whereas those in helical rod domain will lead to cardiomyopathy and muscular dystrophy (80). Therefore, it seems unlikely that the global damage of lamin A/C by HSCT/TBI will directly cause HSCT-PL. However, considering that LMNA mutations for lipodystrophy are supposed to cause altered chromatin organization and changes in downstream gene expression (81), it is tempting that LMNA mutations and HSCT/TBI will result in similar pathological situations.

Although the precise mechanism behind Dunnigan phenotype must wait for future study, embryological difference between adipocytes in head and neck region and those in other sites may be relevant. Whereas craniofacial adipocytes are of ectodermal origin and develop from neural crest stem cells, other adipocyte, both subcutaneous and visceral, are of mesoderm origin (48, 82). In addition, more embryological diversity exists among non-craniofacial adipocytes: whereas visceral adipocytes in epidydimal or retroperitoneal regions derive from paraxial mesoderm, visceral adipocytes in mesenteric and peri-ovarian regions derive from posterior lateral plate mesoderm. Also, most subcutaneous adipocytes develop from posterior lateral plate mesoderm (48). Although those insights are obtained mostly from animal studies, embryological differences may be responsible for developing Dunnigan phenotype.

Next, functional diversity among adipocytes may account for the depot-specific reactions against the detrimental insults such as TBI. Recent studies including single cell analysis has revealed substantial diversity both in adipocytes and APC, and the close correlation between adipocyte function and its surrounding environment (48, 83,84,85,86). For example, a study utilizing single-cell spatial transcriptomics suggested that the functional diversity in APC is closely related to the surrounding local microenvironment, including the proximity with blood vessels, fibrous component, and macrophage (83). In this study, it was found that the mature adipocytes also can be divided into three subpopulations according to their gene expression profiles (83). As stated above, whereas APCs mainly differentiate into mature adipocytes via preadipocytes, they can also differentiate into anti-adipogenic progenitors which will suppress adipogenic differentiation of other APCs (48, 49). All above findings suggest that adipocytes are more diverse even among a single depot. Therefore, the heterogeneity between each adipose depot must be more profound. The pathological basis for Dunnigan phenotype will be elucidated, at least in part, from the functional diversity in adipose tissue between face/neck and buttock/extremities.

3) Predisposition for malignancy and long-term prognosis

CCS following HSCT is already at high risk for subsequent cancer occurrence, namely, second neoplasm (87). On the other hand, there is a concern for developing malignancy, such as lymphoma or head and neck cancers, in patients with FPLD (88, 89). Ancona reported a patient with typical HSCT-PL, who developed breast cancer and died at 29 yr of age (22). Therefore, in the patients with HSCT-PL, surveillance for tumor development should be more cautious than that in CCS without lipodystrophy.

Little is known about long-term prognosis, including mortality and adult-onset complications in HSCT-PL. There is a concern for premature mortality in FPLD, mainly owing to cardiovascular events (35, 36). Because the patients with HSCT-PL must suffer from a wide variety of late complications other than metabolic issues (1,2,3,4,5), invention of innovative treatment, which is discussed in the next section, is an urgent issue.

Future perspective

In mice model of generalized lipodystrophy, adipose transplantation exerted some metabolic benefits (90). No such attempts have been applied to partial lipodystrophy both in mice and humans. Considering the above-mentioned diversity of adipose tissue both in embryological and functional aspects, simplistic method of adipose transplantation may not work in HSCT-PL which is a complex mixture of lipoatrophy and lipohypertrophy. However, transplantation-related therapy may be anticipated, because it is the most realistic way to provide APC pool.

In addition to TZD stated above, other pharmacies have a potency to enhance adipogenesis, such as β-blockers, antidepressants, and antipsychotics (91). Recently, the interaction between bone and adipose tissue has been highlighted, and sclerostin, one of the osteokines produced by osteocytes, has been shown to favor adipocyte differentiation (92). In parallel with the accumulation of the utility of metreleptin and SGLT2 inhibitors, invention of other drugs for HSCT-PL may be anticipated. In fact, for FPLD, some clinical trials have been conducted or are currently underway, which utilizes novel medicines such as vupanorsen (antisense angiopoietin-like 3 protein [ANGPTL3] inhibitor) (93), baricitimab (JAK1/JAK2 inhibitor) (94), and others (summarized in ref. 30). Furthermore, pharmacological interventions for senescence, namely senolytics, may have a role (74) from the perspective connecting HSCT-PL and the accelerated senescence of APC.

As described in the beginning of this Review, CCS following HSCT are at high risk for developing a variety of metabolic dysfunctions, regardless the presence/absence of lipodystrophy (3, 6,7,8,9). Because all the patients with HSCT-PL suffer from metabolic disorder(s) by definition, often with quite severity, it is tempting to regard HSCT-PL as a severest end of the metabolic dysfunctions caused by HSCT. In other words, when underlying pathology is severe enough, HSCT-PL will develop, whereas if the pathology is modest, metabolic dysfunction without definite lipodystrophic phenotype may develop. The possible link between HSCT-PL and an accelerated senescence in APC may lead to broader concept that HSCT-PL is one of the severest complications that occurs in CCS, because CCS, including adult cancer survivors, are associated with accelerated senescence process (58,59,60). From this point of view, research focusing on cellular senescence in HSCT-PL may be of significance.

Conclusion

HSCT-PL seems to be accepted as one of a definite cause for developing APL. Although its precise pathophysiology still remains to be elucidated, HSCT-PL looks like an accelerated senescence process. Female predominance and possible ethnic difference are ascertained. Metreleptin is one of the promising treatment options, and the accumulation of its treatment results are warranted.

Conflict of interests

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data Availability

The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.