Phenotypic diversity of frontotemporal lobar degeneration in two novel GRN variants from Colombia

Abstract

INTRODUCTION

Pathogenic progranulin (GRN) variants are among the main genetic causes of frontotemporal lobar degeneration (FTLD). These variants have been predominantly reported in European cohorts, but their characterization in Latin America remains scarce. We describe two Colombian cases with novel GRN variants with amnestic and semantic syndromes leading to an initial diagnosis of Alzheimer's disease (AD).

METHODS

We conducted clinical, neuropsychological, neuroimaging, and genetic analysis. Biomarkers were included for one case.

RESULTS

At 42 years old, Case 1 presented a predominant amnestic profile and carried the GRN c.21G > A (p.Trp7*) variant. Case 2 debuted with a semantic impairment at 62 years old and was a carrier of GRN c.1098T > A (p.Cys366*) variant. Brain imaging revealed asymmetric temporal atrophy, and biomarkers supported diagnosis of FTLD.

DISCUSSION

GRN variants can mimic early‐onset AD. An integrative approach including serial clinical, genetic, brain imaging, and biomarker analysis are essential for diagnosing Amnestic variants of FTLD in admixed genetic populations.

Highlights

• Two cases with previously unreported progranulin (GRN) pathogenic variants are shown.

• Both cases are from Latin America, a population understudied for frontotemporal lobar degeration (FTLD).

• Cases present atypical cognitive syndromes at onset challenging diagnostic approach.

• We proposed the use of the term “amnestic variant of FTLD” in clinical practice.

1. BACKGROUND

Frontotemporal lobar degeneration (FTLD) is a group of neurodegenerative conditions characterized by behavioral, language, and motor impairments impacting activities of daily living (ADL). ¹ , ² , ³ Although most cases of FTLD are sporadic, studies report that up to 20%–25% of patients may carry a pathogenic genetic variant, most commonly in C9orf72, GRN, MAPT, FUS, or TARDBP. ⁴ , ⁵

Heterozygous loss‐of‐function variants in the Progranulin gene, GRN (NM_2087), have been recognized as one of the major genetic causes of FTLD associated with TAR DNA‐binding protein 43 (TDP‑43) inclusions (FTLD‐TDP43). ⁴ Most patients present with behavioral variant of frontotemporal dementia (bvFTD), but uncommon presentations, such as pure amnestic syndromes and parkinsonism have been described. ⁶

These atypical cognitive syndromes associated with GRN variants have been almost exclusively reported in patients of European and North‐American origin. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Although the literature from Latin America remains limited, pathogenic GRN variants with diverse clinical phenotypes have been identified in population‐based studies from Argentina, ¹³ , ¹⁴ Brazil, ¹⁵ the Caribbean, ¹⁶ , ¹⁷ Chile, Colombia, and Peru. ¹⁸

Despite major advances to study the FTLD spectrum in Latin America, the current population‐based studies remain unable to provide in‐depth characterization of the clinical, cognitive, and neuroimaging features of genetic FTLD in the region. Identifying clinical patterns to differentiate GRN carriers from other neurodegenerative conditions is especially important in Latin America where FTLD is usually misdiagnosed due to structural and administrative limitations. ¹⁹ , ²⁰ , ²¹ This case series present a clinical, neuropsychological, neuroimaging, and biomarker analysis of two Colombian cases with two novel pathogenic variants: GRN (NM_002087.4) c.21G > A p.Trp7* and c.1098T > A p.Cys366*. Both patients debuted with atypical early amnestic and semantic syndromes, but thanks to longitudinal assessment and genetic testing were able finally diagnosed with FTLD.

2. METHODS

2.1. Clinical and neuropsychological assessment

Patients were evaluated within the Multipartner Consortium to Expand Dementia Research in Latin America (ReD‐Lat) study in two centers from Colombia. ¹⁸ Written informed consent was obtained for participation and disclosure of clinical and genetic findings. All procedures were approved by the Institutional Review Boards (FWA00028864 and FWA00001113). Patients underwent comprehensive assessments including a semi‐structured interview, complete clinical and neurological examination, and neuropsychological assessment. The diagnosis was made using the current criteria for bvFTD and svPPA, ³ , ²² respectively. Diagnoses of AD dementia syndromes were considered according to the International Working Group criteria. ²³

2.2. Neuroimaging analysis

The voxel‐based morphometric (VBM) analysis of structural data was performed as described by Ossenkoppele et al. 2015. ²⁴

2.3. Biomarker and genetic analyses

Blood and cerebrospinal fluid (CSF) samples were acquired following international standards for obtention, processing, and storage. DNA extraction, whole genome sequencing (WGS) and variant curation were performed as described in Acosta‐Uribe et al 2025. ²² GRN variants were validated though sanger  sequencing.

FIGURE 1.

Dot plots comparing neuropsychological performance using normalized z‐scores according to age and education. Each test was grouped according to cognitive domain for case 1 at ages 58 and 62. Dashed lines represent the z‐score cutoff for cognitive impairment according to DSM‐5 criteria. BNT, Boston Naming Test; CP, CERAD constructional praxis; Ex.F, executive function; INECO‐FST, Instituto Neurologico‐frontal Screening Test; Lg, language; MCT, Memory Capacity Test; MIS, Memory Impairment Screening; MMSE, Mini Mental State Examination test; PS, processing speed; RCFT, Rey‐Osterrieth Complex Figure; TMT‐A, Trail Making Test‐A; WAIS‐III, Wechsler Adult Intelligence Scale, Spanish‐III; WCST, Wisconsin Card Sorting Test; Vs, visuo‐constructive skills.

For Case 1, CSF biomarker concentrations of total tau, phosphorylated tau (pTau‐181), and amyloid‐β42 (Aβ42) were quantified using a Lumipulse® electrochemiluminescent immunoassay.

3. CASE REPORT

3.1. Case 1

A 58‐year‐old admixed man from the central region of Colombia, with 16 years of formal education and lifelong career as a company treasurer, presented with memory and behavioral complaints that interfered with his ADL. His family history was notable for late‐onset dementia in his mother and multiple sclerosis in a deceased brother, both with unknown genetic status (Figure S1). At age 42, the participant noticed subtle memory difficulties that progressed over the following decade and began to interfere with ADL at the age of 54. These symptoms led to difficulties at work, which he was able to partially compensate for. His wife also reported episodes of confabulation and emergence of unusual habits, such as highlighting articles in the newspaper and recording daily news in a notebook.

At age 58, he was demoted at work due to forgetfulness, prompting neurologic consultation. After ruling out secondary causes, he was diagnosed with dementia. Neuropsychology testing showed major impairment in episodic memory and verbal fluency, along with mild behavioral symptoms (Figure 1). Cognitive assessment showed reduced information storage during controlled encoding that improved slightly with retrieval cues (Figure 1). Language, visuo‐constructional abilities, and attentional were intact. An amnestic hippocampal syndrome consistent with typical AD was diagnosed, and pharmacological treatment was initiated.

RESEARCH IN CONTEXT

1. Systematic review: A literature review using PubMed and Google Scholar was done. We identify several cases of pathogenic progranulin (GRN) variants with amnestic syndrome and current GRN reported variants in Latin America.

2. Interpretation: Current reports have not considered GRN amnestic cases in Latin America, leaving a gap in the clinical diversity of GRN and genetic frontotemporal lobar degeneration (FTLD) in our region.

3. Future directions: This study shows a comprehensive longitudinal characterization of two novel GRN mutations in Colombia, highlighting the wide manifestations of FTLD in Latin America. Future studies should expect amnestic variants of FTLD in pathogenic GRN cohorts. Also, studies should consider the impact of admixed ancestry on clinical manifestations of the Latin America population.

At age 60, a brain magnetic resonance imaging (MRI) revealed bilateral of the medial temporal lobe atrophy, including the head of the hippocampus. The atrophy was more prominent on the right side, with a Scheltens score of I–II, and was associated with enlargement of the right temporal horn of the lateral ventricle (Figure 2A). Few small, bilateral, non‐confluent white hyperintensities in frontal white matter were also identified on fluid attenuated inversion recovery (FLAIR) sequence (Figure S2).

FIGURE 2.

Neuroimaging findings of Case 1 (A) MRI T1 and T2 sequences obtained at age 60 showing moderate asymmetric atrophy of the right temporal lobe. An increase in ventricular space is visible in the right temporal lobe on the coronal T2 view, with apparent right hippocampal atrophy. (B) F[18]‐FDG‐PET imaging obtained at age 62 showing non‐normalized radiotracer uptake with asymmetric hypometabolism in the right temporoparietal structures and the right caudate nucleus. FDG, fluorodeoxyglucose; MRI, magnetic resonance imaging; PET, positron emission tomography.

Later, at age 62, he exhibited increased irritability, stereotyped behaviors, ritualistic behaviors, and visual hallucinations (Table S1). Memory concerns still prevailed but remained relatively stable. A new neuropsychological assessment revealed severe anterograde amnesia, spatial disorientation, and the emergence of disinhibition symptoms, represented by difficulties with motor inhibitory tasks and altered working memory during the INECO Frontal Screening test.

Given the previous inconsistencies with a diagnosis of AD, a fluorodeoxyglucose‐positron emission tomography (FDG‐PET) scan was obtained. It revealed greater right temporal atrophy compared to his previous MRI and also showed generalized hypometabolism in the right hemisphere, most pronounced in the frontotemporal region. Hypometabolism also involved the right caudate nucleus and the right inferior parietal region (Figure 2B). CSF AD biomarkers were also ordered, showing positive Aβ1‐42 levels, negative total‐tau and negative p‐tau 181 (Table S1, Case 1). Results pointed towards probable AD neuropathological change. However, negative tau biomarkers in the context of overt dementia were against AD as the primary neuropathological contributor.

Whole genome sequencing of the patient identified a novel variant in the GRN gene: NM_002087.4:c.21G > A (p.Trp7*), which was later confirmed by Sanger sequencing. The GRN p.Trp7* was determined to be pathogenic according to the American College of Medical Genetics and Genomics (ACMG) guidelines (PVS1 criterion). ²⁵ No pathogenic variants in the APP, PSEN1, or PSEN2 genes were identified, and apolipoprotein E (APOE) genotype was ε3/ε3 (Table S1). We conclude that the patient had a probable FTLD‐TDP43 pathology with an amnestic syndrome presentation.

3.2. Case 2

A 70‐year‐old admixed woman from the central region of Colombia, with 12 years of formal education and a background in accounting‐related work, presented with prominent short‐term memory and semantic difficulties around age 62 (Table S1). Her family history was notable for svPPA, in her brother, early‐onset amnestic impairment in her sister, and dementia in her maternal grandfather (Figure S3).

At age 67, she was invited to participate in a screening clinical assessment based on the previous identification of svPPA in her brother. Physical and neurological examinations were unremarkable, and laboratory results ruled out secondary causes of dementia. Her neuropsychological assessment showed reduced ability to learn and store information in verbal memory tasks, along with less efficient retrieval of semantic information and mild impairment in executive function (Figure 3). These cognitive outcomes did not impact ADL. Based on this, she was considered to have amnestic multidomain mild cognitive impairment (MCI), consistent with an early stage of AD (Table S1). A follow‐up examination at age 69 showed similar neuropsychological results to the first visit, with mild worsening in processing speed and slight improvement on frontal screening tests (Figure 3).

FIGURE 3.

z‐Normalized neuropsychological test performance according to age and education for case 2 at age 67, 69, and 70. Dashed lines represent z‐score cutoff for cognitive impairment according to DSM‐5 criteria. DSM‐5, Diagnostic and Statistical Manual of Mental Disorders, Fifth Revision; Ex.F, executive function; FCSRT, Free and Cued Selective Reminding Test; INECO‐FST, Instituto Neurologico‐frontal Screening Test; Lg, language; MMSE, Mini Mental State Examination test; PS, processing speed; RCFT, Rey‐Osterrieth Complex Figure; TMT‐A, Trail Making Test‐A; Vs, visuo‐constructive skills.

At age 70, her relatives reported subtle changes in recalling names, finding words, understanding spoken language, and apathy, which were impacting her ADL. Her neuropsychological assessment showed subtle worsening in semantic, language, attentional, and executive functions (Figure 3). Memory testing indicated adequate encoding, slightly reduced consolidation of verbal information, and limited long‐term recall that remained unchanged compared to her two previous assessments. Her language was fluent with intact comprehension and repetition, but mild anomia was observed. Semantic categorization tasks were performed better on general categories (83%) than on specific ones (73%), indicating reduced access to detailed semantic features across both visual and verbal modalities (Figure S4A–C). She also showed impairments in rule‐based and arbitrary Spanish spelling (Figure S4D), as well as difficulties identifying real and made‐up objects (Figure S4E). This was consistent with a clear semantic impairment that would not be expected in an AD diagnosis.

Brain MRI revealed global atrophy with a left‐predominant pattern. The atrophy was most pronounced in the left medial temporal lobe, including the hippocampal formation, as well as the left inferior temporal lobe (Figure 4A). This asymmetry was validated using VBM, which showed a clear left‐sided hippocampal and medial temporal lobe involvement, with lateral temporal atrophy compared with matched healthy control templates paired by sex, age, and geographical region (Figure 4B).

FIGURE 4.

Neuroimaging findings of case 2. (A) MRI T1 showing global atrophy with greater involvement of the left temporal lobe. Left hippocampal atrophy was also apparent. (B) Voxelwise W‐scores were generated by comparing case 2 modulated grey matter + white matter tissue probability maps against a normative model constructed from age, sex, and scanner‐matched healthy controls. Negative values (W < 2.0) indicate tissue loss beyond the expected distribution, after accounting for nuisance covariates. Regional clusters of atrophy are evident in the left temporal middle and superior temporal regions, including hippocampus. MRI, magnetic resonance imaging.

WGS identified the GRN variant NM_002087.4:c.1098T > A (p.Cys366*) variant, which was validated through Sanger sequencing (Figure S5). The GRN p.Cys366* was determined to be pathogenic according to ACMG guidelines (PVS1 criterion). ²⁵ No pathogenic variants in the APP, PSEN1, or PSEN2 genes were identified, and APOE genotype was ε3/ε3 (Table S1). These results confirmed an FTLD diagnosis presenting as svPPA.

4. DISCUSSION

Most GRN‐related FTLD cases in the literature are focused on bvFTD and dominated by reports from European, North American, and Asian cohorts, ²⁶ leaving admixed populations, like those in Latin America, understudied. The novel pathogenic GRN variants reported in this study (p.Trp7* and p.Cys366*), show evidence of clinical heterogeneity in admixed populations. Case 1 (GRN p.Trp7*) presented with a pure amnestic syndrome with late behavioral impairment, whereas Case 2 (GRN p.Cys366*) initially presented with amnestic MCI accompanied by subtle semantic deficits that later evolved into svPPA. These early findings led to an incorrect diagnosis of AD in both cases, which was later revised following integrative multidisciplinary follow‐up, resulting in a FTLD‐GRN diagnostic.

Amnestic presentations are relatively common in FTLD‐GRN. In a large population‐based study, probable AD was diagnosed in 14% of patients with GRN mutations; episodic memory impairment was observed in in 72.7%, and a hippocampal amnestic syndrome resembling AD was present in 18.2% of cases. ²⁷ These findings align with previous reports of GRN‐related amnestic syndromes characterized by early memory deficits that are clinically indistinguishable from AD, followed by atypical cognitive progression involving language or behavioral changes (i.e., stereotyped or ritualistic behaviors). ¹⁰ , ²⁸ , ²⁹

However, the presence of AD co‐pathology in many of these cases cannot be excluded. Indeed, some GRN‐related amnestic cases have shown positive CSF biomarkers consistent with AD co‐pathology. ¹⁰ , ³⁰ , ³¹ For example, Perry et al ¹⁰ reported a case similar to our first patient, where an individual initially presented with episodic memory concerns that later evolved into an FTLD diagnosis. This patient had a positive Pittsburgh Compound B (PiB)‐PET tracer and carried the GRN c.1263_1264insGAAGCGAG variant, suggesting TDP‐43 pathology as the primary contributor. However, the phenotypic contribution of TDP‐43 pathology remained unclear due to the presence of an APOE ε4 allele and the absence of postmortem neuropathological studies. Together, these findings raise the possibility that co‐occurring AD‐related neuropathological changes contributed to the observed clinical phenotypes. Alternatively, co‐pathology may exacerbate atypical amnestic presentation in these cases, potentially reflecting convergent neuropathological mechanisms involving APOE and GRN. ³²

Although we were unable to definitively exclude other causes of dementia in the absence of plasma progranulin measurements and neuropathological analyses, we did not identify any pathogenic variants in APP, PSEN1, PSEN2, nor the risk allele APOE ε4 that would explain a pure AD diagnosis. Additionally, one of our patients had negative tau biomarkers (Table S1), reinforcing the hypothesis that GRN haploinsufficiency was driving the cognitive decline in these cases.

Our cases met some of the core criteria for possible limbic‐predominant age‐related TDP‐43 encephalopathy (LATE), such as predominant amnestic syndrome with relative preservation of other cognitive domains. ³³ Nonetheless, the presence of pathogenic variants in GRN excludes LATE diagnosis. ³⁴ Consequently, these patients met the criteria of low certainty for limbic‐predominant amnestic neurodegenerative syndrome based on their clinical and neuroimaging findings. ³⁵

Recent advances on the clinical‐biological definition of AD have highlighted the heterogenous clinical presentations of neuropathology‐confirmed AD, including the behavioral/dysexecutive variant. ²⁴ , ³⁶ The clinical presentation observed in p.Trp7* and p.Cys366* GRN carriers fall within the proposed FTD‐AD‐LATE clinical spectrum ²³ in which GRN‐mediated TDP‐43 pathology can produce an early‐onset anterograde amnestic syndrome. We propose the term “amnestic variant of FTLD” to describe pure hippocampal amnestic syndromes occurring in patients with confirmed FTLD pathology or pathogenic variants in FTLD‐associated genes, in the absence of alternative causes of dementia. This terminology reflects the underlying biological basis of cognitive impairment and aligns with the recommendations of the Dementia Nomenclature Initiative. ³⁷ Although distinguishing an “amnestic variant of FTLD” from LATE might be challenging in clinical practice, neuropathological studies have demonstrated distinguishing features between the two entities, ³⁸ supporting the partial independence of their pathophysiological mechanisms.

This partial independence of pathological mechanisms may be reflected in distinct neuroimaging patterns. In the present cases, neuroimaging revealed asymmetric temporal lobe atrophy, contrasting with the symmetric hippocampal atrophy typically observed in AD ²⁴ and the predominantly symmetric frontal lobe atrophy reported in other genetic forms of FTLD. ²⁴ , ³⁹ In LATE, neuroimaging is characterized by disproportionate hippocampal atrophy relative to the rest of the brain. ³³ , ⁴⁰ Although Case 1 initially showed mild right hippocampal atrophy (Scheltens I–II), progressive temporal lobe atrophy rapidly exceeded hippocampal involvement, further supporting a distinct neurodegenerative pattern. Case 2 exhibited asymmetric left‐sided atrophy predominantly affecting the anterior temporal lobe, consistent with previous GRN‐svPPA descriptions. ³⁹ , ⁴¹ Notably, both cases showed early atrophy of the hippocampal formation, which differs from patterns reported in other GRN pathogenic variants. ¹¹ , ³⁹ , ⁴² , ⁴³ These findings suggests that early asymmetric hippocampal atrophy may be an early indicator of GRN‐related amnestic presentations.

Genetic admixture in Latin America may also contribute to atypical presentation of FTLD. In a previous study of neurodegeneration in Colombia, patients who carrying pathogenic variants in amyotrophic lateral sclerosis‐related genes, including SOD1, on Amerindian haplotypes often presented FTLD without motor neuron involvement. ¹⁴ Similarly, case reports from a Caribbean cohort described amnestic symptoms in carriers of the GRN risk variant 3′UTR c.*78C > T (rs5848). ⁴⁴ Interestingly, the same risk‐conferring variant has been associated with increased risk for LATE in North American cohorts. ⁴⁴ , ⁴⁵ Together, these findings suggest that genetic background and ancestry‐related modifiers may influence the phenotypic expression of GRN‐related neurodegeneration.

These two cases demonstrate how pathogenic variants in GRN can present non‐canonical clinical syndromes, complicating diagnostic classification within the FTD‐AD‐LATE spectrum. We propose the term ‘amnestic variant of FTLD’ to account for the prodromal amnestic syndrome, as described in these two cases with pathogenic GRN variants. Together, these findings highlight that accurate diagnosis of FTLD requires longitudinal, multimodal assessment integrating clinical, cognitive, neuroimaging, biomarker, and genetic data, especially when early clinical features are atypical.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

CONSENT STATEMENT

Both individuals included in this paper received and signed an informed consent approved by the institutional review boards of Pontificia Universidad Javeriana—Hospital Universitario San Ignacio (FWA00001113) and Instituto de Investigaciones Médicas—Facultad de Medicina, Universidad de Antioquia (FWA00028864). All procedures in this project followed the Helsinki declaration and adhered to the international and national normative for research in humans including the Colombian Resolutions 8430/1993 and 2378/2008. Both participants and caregivers authorized for the publication of the anonymized clinical data under informed consents.

Supporting information

Supporting Information: alz71475‐sup‐0001‐SuppMat.docx

Supporting Information: alz71475‐sup‐0002‐SuppMat.pdf