Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Association between Plasma Homocysteine Levels and Neuronal Injury in HIV Infection

  • Erika Ahlgren ,

    Affiliation Department of Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

  • Lars Hagberg,

    Affiliation Department of Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

  • Dietmar Fuchs,

    Affiliation Division of Biological Chemistry, Biocenter, Innsbruck Medical University, Innsbruck, Austria

  • Lars-Magnus Andersson,

    Affiliation Department of Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

  • Staffan Nilsson,

    Affiliation Mathematical Sciences, Chalmers University of Technology, Gothenburg, Sweden

  • Henrik Zetterberg,

    Affiliations Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Mölndal, Sweden, UCL Institute of Neurology, Queen Square, London WC1N 3BG, United Kingdom

  • Magnus Gisslén

    Affiliation Department of Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

Association between Plasma Homocysteine Levels and Neuronal Injury in HIV Infection

  • Erika Ahlgren, 
  • Lars Hagberg, 
  • Dietmar Fuchs, 
  • Lars-Magnus Andersson, 
  • Staffan Nilsson, 
  • Henrik Zetterberg, 
  • Magnus Gisslén



To investigate the role of homocysteine in neuronal injury in HIV infection.


Using a cross-sectional design and archived samples, we compared concentrations of plasma homocysteine and cerebrospinal fluid (CSF) neurofilament light protein (NFL), a sensitive marker of neuronal injury, in 83 HIV-1-infected subjects without antiretroviral treatment. We also analyzed plasma vitamin B12, serum folate, CSF, and plasma HIV RNA, the immune activation marker neopterin in CSF and serum, and albumin ratio as a marker of blood-brain barrier integrity. Twenty-two subjects provided a second sample median of 12.5 months after antiretroviral treatment initiation.


A significant correlation was found between plasma homocysteine and CSF NFL concentrations in untreated individuals (r = 0.52, p < 0.0001). As expected, there was a significant inverse correlation between homocysteine and B12 (r = –0.41, p < 0.001) and folate (r = –0.40, p = < 0.001) levels. In a multiple linear regression analysis homocysteine stood out as an independent predictor of CSF NFL in HIV-1-infected individuals. The correlation of plasma homocysteine and CSF NFL was also present in the group receiving antiretroviral therapy (r = 0.51, p = 0.016).


A correlation between plasma homocysteine and axonal injury, as measured by CSF NFL, was found in both untreated and treated HIV. While this study is not able to prove a causal link, homocysteine and functional B12/folate deficiency appear to play a role in neural injury in HIV-infected individuals.


Prior to the discovery of effective antiretroviral therapy (ART), HIV-infected individuals had about a 30% overall risk of developing HIV-related dementia (HAD) [1]. Although the picture today is different, HIV diagnosis is to some extent still associated with HIV-associated neurocognitive disorders (HAND). Symptoms of HAND are common in untreated HIV-positive individuals and frequently reported by individuals on suppressive ART [2]. Recent studies have noted increased levels of cerebrospinal fluid (CSF) neurofilament light protein (NFL) in HIV-infected individuals, which is interpreted as a sign of ongoing neuronal injury. Elevated levels of CSF NFL have not only been reported in individuals with neurological symptoms, but also in some individuals with and without ART who are asymptomatic with regard to neurological symptoms [3,4].

Associations between elevated plasma homocysteine levels and cognitive impairment in HIV-negative individuals have been the topic of many studies. Data suggest an association between elevated levels of homocysteine and diseases of cognitive impairment such as Alzheimer’s [5,6]. However, in the case of HIV-infected individuals, data on homocysteine levels in the context of CNS injury and neurocognition are rare. The metabolism of homocysteine is dependent on folate and vitamins B12 and B6, and therefore elevated levels of homocysteine are an indicator of B12 and/or folate deficiency (Fig 1) [7]. This study investigates the possible association between homocysteine and neuronal injury in HIV-1-infected individuals.

Fig 1. Homocysteine metabolism.

Abbreviations: SAM, S-adenosyl-methionine; SAH, S-adenosyl-homocysteine; MTHF, methylenetetrahydrofolate; THF, tetrahydrofolate.


Study design

A retrospective selection was made of samples collected in a prospective research program, as previously described [8]. All samples were collected between 1999 and 2014 at the Department of Infectious Diseases at Sahlgrenska University Hospital, Gothenburg, Sweden. Study participants were randomly chosen from individuals in the cohort who met the inclusion criteria, i.e. untreated HIV-infection, over 18 years old, and no opportunistic CNS complications. Plasma and CSF specimens from 83 HIV-infected untreated individuals were analyzed (Table 1). Another plasma and CSF sample was obtained at a median of 12 months (range 10.8 to 27.1) in 22 subjects on ART. Subjects without signs of neurological or cognitive impairment at clinical examination were defined as neuroasymptomatic. Fifty-three of those had an asymptomatic HIV-infection (Center for Disease Control and Prevention [CDC] classification A1-A3) [9]. Twenty-two patients fulfilled the criteria for AIDS (CDC classification C2-C3), 13 due to pneumocystis pneumonia (PCP), four with tuberculosis, two with wasting syndrome, and one each with Kaposi’s sarcoma, extracerebral lymphoma, and salmonella sepsis. Five patients had symptomatic HIV-infections not categorized as AIDS (CDC classification B2-B3), all due to oral candidiasis.

The untreated neuroasymptomatic individuals were divided into four subgroups according to CD4+ T-cell count (T-cells/μL): CD4+ > 350 (n = 21); CD4+ 200–349 (n = 20); CD4+ 50–199 (n = 22); and CD4+ < 50 (n = 17), based on earlier findings of higher CSF-NFL levels in untreated subjects with lower CD4 cell counts [3]. Three additional patients with HAD were tested but not included in the statistical analyses. All predictors were selected based on earlier findings of relevance to CNS HIV infection [3,10].

Laboratory assays

All samples were frozen to –70°C within an hour of sampling and stored until analysis. CSF samples were centrifuged to remove cells and aliquoted before freezing. Levels of NFL in CSF were determined using a commercial ELISA, according to the manufacturer’s instructions (Uman Diagnostics, Umeå, Sweden). For neopterin a commercially available immunoassay (BRAHMS, Hennigsdorf, Germany) was used. CD4+ count, CSF, and blood albumin levels were analyzed according to local laboratory standards. CSF:blood albumin ratio was calculated as a measure of blood-brain barrier dysfunction.

Plasma vitamin B12 and serum folate were quantified using an electrochemiluminescence immunoassay on a cobas e analyzer (Roche, Penzberg, Germany). Total homocysteine was determined in EDTA plasma by stable isotope dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) using a Quattro micro instrument (Waters Corporation, Milford, MA, USA), essentially as described by Magera et al. [11]. Levels of HIV RNA, both in CSF and plasma, were determined using the Roche Amplicor version 1.5, or Roche COBAS TaqMan assay version 1 or 2 (Hoffman La-Roche, Basel, Switzerland)

Statistical analysis

Correlations were analyzed by Pearson correlation analysis. Paired T-test was used to compare the treatment group before and during treatment. Differences between groups were analyzed with ordinary one-way ANOVA and P values were adjusted for multiple comparisons using the Tukey’s Range test. Predictors of log10 plasma homocysteine and CSF NFL were analyzed by multiple linear regression analysis with forward selection. In the analyses CD4+ count, CSF:blood albumin ratio, plasma and CSF HIV RNA, serum and CSF neopterin, CSF NFL, plasma homocysteine, plasma vitamin B12, and serum folate were log transformed to reduce skewness. P < 0.05 was considered significant.

All statistical tests were performed with SPSS version 21 (IBM SPSS Statistics, Armonk, NY, USA) or Prism version 6.0 (Graphpad Software Inc., La Jolla, CA, USA).

Ethics approval and informed consent

The entire study and the procedure used to assure the anonymity of the data was approved by the Research Ethics Committee of the University of Gothenburg, in accordance with the Helsinki Declaration of 1975, as revised in 2000. All participants gave their informed consent in writing. All patient information was anonymized prior to analysis by study Principal investigator (MG) at the Department of Infectious Diseases, at Sahlgrenska University Hospital.


Descriptive statistics

Among the 80 untreated neuroasymptomatic subjects, 20 had moderate hyperhomocysteinemia (> 15 μmol/L). Twenty had elevated levels of CSF NFL, compared to age-dependent laboratory norms. B12 levels were below 140 pmol/L in 2 subjects, and below 250 pmol/L in 32. One individual had a high serum folate concentration. Of the three individuals with HAD, two had hyperhomocysteinemia and two had CD4+ counts < 200. All three had very high levels of CSF HIV RNA and CSF NFL.

Group comparisons

Comparison of the subgroups of untreated neuroasymptomatic subjects showed a significant difference in homocysteine levels between the groups (p = 0.0002). The group with CD4+ count < 50 had homocysteine levels significantly higher than the other groups (Fig 2a). Differences between groups were also detected in levels of CSF NFL (p = 0.0001). The CD4+ < 50 group had significantly higher levels than the groups with CD4+ > 350 and CD4+ 200–349. The CD4+ 50–199 group had higher levels compared to the CD4+ > 350 group (Fig 2b). No significant difference between groups was found in plasma B12 or serum folate levels (Fig 2c and 2d).

Fig 2. A-F. Levels of homocysteine, NFL, vitamin B12, and folate. Correlations of homocysteine and CSF NFL.

A-D: Levels of homocysteine, NFL, vitamin B12, and folate in a cohort of HIV-infected individuals divided into groups. Neuroasymptomatic individuals are divided according to CD4+ T-Cell levels. Individuals with HAD and those on ART are presented as individual groups. * Adjusted p values. E: Correlation of Log P homocysteine and Log CSF NFL in 80 untreated, neuroasymptomatic HIV-infected individuals. F: Correlation of Log P homocysteine and Log CSF NFL in 22 neuroasymptomatic HIV-infected individuals on ART. Abbreviations: P, plasma; HAD, HIV associated dementia; ART, antiretroviral therapy; CSF, cerebrospinal fluid; NFL, neurofilament light protein.

Predictors of CSF NFL and homocysteine

We found a significant correlation (r = 0.51, p < 0.0001) between the plasma level of homocysteine and CSF NFL (Fig 2e). As expected, there was also a significant inverse correlation of homocysteine and B12 (r = –0.36, p = 0.0009), and of homocysteine and folate (r = –0.40, p = 0.0003). Homocysteine was also found to correlate significantly to the CSF:blood albumin ratio (r = 0.25, p = 0.023), CD4+ count (r = –0.41, p = 0.0002), age (r = 0.40, p = 0.0002), and treatment with trimethoprim-sulfamethoxazole (r = 0.355, p = 0.001). Plasma homocysteine correlated to serum neopterin (r = 0.34, p = 0.002), but no correlation was found between homocysteine and CSF neopterin. No significant correlation was found between CSF NFL and B12 concentrations. A weak correlation was found between serum folate and CSF NFL (r = 0.28, p = 0.012).

Using multiple linear regression analysis age, plasma vitamin B12, serum folate, plasma HIV RNA, and treatment with trimethoprim-sulfamethoxazole were found to predict plasma homocysteine; however, serum neopterin and CD4+ count did not (Table 2). Plasma homocysteine, CSF neopterin, age, plasma HIV RNA, and CD4+ count stood out as independent predictors of CSF NFL, with adjusted estimates in the multivariate analysis. By contrast, CSF:blood albumin ratio, CSF viral load, serum neopterin, plasma vitamin B12, and serum folate were not found to be significant predictors (Table 3).

CSF NFL and homocysteine correlation after initiation of ART

The correlation of plasma homocysteine and CSF NFL was still present in the group on ART (Fig 2f). No significant differences in levels of plasma homocysteine or CSF NFL were found in the ART group before compared to during ART.


Our results show an independent correlation of plasma homocysteine and CSF NFL concentrations in neuroasymptomatic HIV-infected individuals, both untreated and those on ART. To the best of our knowledge, this is the first report that shows an association in any disease between NFL, a sensitive biomarker of neuronal injury, and plasma homocysteine.

Earlier studies have shown the prevalence of hyperhomocysteinemia in untreated HIV-infected individuals to be between 20% and 35% [12,13]. This is similar to our finding of 25%. Homocysteine levels increase with B12 vitamin and folate deficiency and are also influenced by renal function, age, and treatment with certain medications, e.g. trimethoprim. None of the subjects in our study had impaired renal function, and retrospective analysis of medical records showed that none were taking B vitamins. However, we discovered elevated levels of vitamin B12 in plasma in 14 samples, suggesting the possibility of B12 vitamin supplementation that was not recorded in patient files.

Trimethoprim treatment increased levels of homocysteine in a small number of healthy men [14]. However, no changes in homocysteine were seen in HIV-positive individuals on prophylactic doses of trimethoprim-sulfamethoxazole [15]. Seven individuals in our cohort received prophylactic and 13 treatment doses of trimethoprim-sulfamethoxazole. The correlation found between homocysteine and CD4+ cell count in the univariate analysis disappeared in the multivariate analysis. This may in part be explained by trimethoprim-sulfamethoxazole treatment of subjects with pneumocystis jiroveci pneumonia and low CD4+ cell count.

Elevated levels of CSF NFL are nearly always present in individuals with HAD [3,4], and increased levels can already be found 1 to 2 years before the development of overt dementia symptoms [8]. Consequently CSF NFL has been proposed as a predictive marker of HIV-associated neurocognitive disease before clinical symptoms appear [4,8]. In untreated neuroasymptomatic, HIV-infected individuals, markedly elevated CSF NFL levels signifying ongoing axonal injury are mainly found in those with low CD4+ cell counts [3]. Although our findings do not prove a causal link, they suggest homocysteine and functional B12/folate deficiency may be involved in the pathogenesis of CNS neural injury in HIV infection. In agreement with earlier findings, CSF NFL was higher in subjects with low CD4+ cells. CSF NFL also correlated with age and CSF neopterin. Moreover, as expected, correlations existed between age, B12, folate, and homocysteine.

HIV-infected individuals have increased levels of neopterin in CSF and serum as a result of immune activation [10]. Earlier studies of HIV-negative individuals have found a correlation between homocysteine and serum neopterin, implying an association between homocysteine and immune activation [16]. It has been hypothesized that neopterin, through the influence on folate metabolism, inhibits methionine synthase [17]. In addition, nitric oxide from activated macrophages may also inhibit the enzyme [18]. Another proposed explanation of the role of immune activation in hyperhomocysteinemia is through the oxidation of B12 and folate. This could give rise to a functional deficiency that has the potential of both enhancing oxidative stress and perturbing the methylation process [19]. However, we did not find serum neopterin to be an independent predictor of homocysteine in multivariate analysis. In an earlier study of individuals with dementia, treatment with B vitamins reduced the level of homocysteine but did not influence neopterin concentrations [20].

The measurement of B12 is a poor marker of B12 deficiency, whereas homocysteine is more sensitive [21]. This may partly explain the lack of correlation we found between CSF NFL and plasma vitamin B12. The inhibition of crucial enzymes, or the lack of adequate forms of B12/folate, may also explain disturbances in the homocysteine metabolism, even when there are sufficient levels of B12 and folate. The homocysteine lowering effect of B vitamin treatment supports the presence of functional B12 and/or folate deficiency in the HIV-negative population. The limited data available indicates that this is also the case in HIV-positive individuals [22].

Significantly, the correlation between CSF NFL and plasma homocysteine remained present in the group on ART. We found a correlation between homocysteine and HIV RNA levels in serum in untreated HIV, but we did not find a significant change in homocysteine levels before compared to after ART initiation.

Previous studies of the association of B vitamin levels and neurocognitive function in untreated HIV-infected individuals have been inconclusive [2325]. Impaired information processing speed and visuospatial problem solving abilities have been reported in untreated, neuroasymptomatic HIV-infected individuals with B12 deficiencies [23]. A small study reported improvement of neuropathic symptoms in such individuals when given replacement therapy [24].

There are several limitations to our pilot study: it is of retrospective design; the sample size was limited; no HIV-negative control group was included; and neuropsychiatric testing was not performed on all asymptomatic subjects.


We found a correlation between plasma homocysteine and axonal injury, as measured by CSF NFL, in both untreated and treated individuals with HIV. While our study is unable to prove a causal link, homocysteine and functional B12/folate deficiency appear to play a role in neural injury in HIV-infected individuals that is worth exploring.

Supporting Information

S1 Table. Full set of the data included in the analysis.


Author Contributions

Conceived and designed the experiments: MG. Performed the experiments: MG HZ DF LMA LH. Analyzed the data: EA MG SN. Contributed reagents/materials/analysis tools: HZ DF. Wrote the paper: EA MG.


  1. 1. Portegies P, de Gans J, Lange JM, Derix MM, Speelman H, et al. (1989) Declining incidence of AIDS dementia complex after introduction of zidovudine treatment. Bmj 299: 819–821. pmid:2510843
  2. 2. Heaton RK, Clifford DB, Franklin DR Jr., Woods SP, Ake C, et al. (2010) HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER Study. Neurology 75: 2087–2096. pmid:21135382
  3. 3. Jessen Krut J, Mellberg T, Price RW, Hagberg L, Fuchs D, et al. (2014) Biomarker Evidence of Axonal Injury in Neuroasymptomatic HIV-1 Patients. PLoS One 9: e88591. pmid:24523921
  4. 4. Abdulle S, Mellgren A, Brew BJ, Cinque P, Hagberg L, et al. (2007) CSF neurofilament protein (NFL)—a marker of active HIV-related neurodegeneration. J Neurol 254: 1026–1032. pmid:17420923
  5. 5. Wald DS, Kasturiratne A, Simmonds M (2011) Serum homocysteine and dementia: meta-analysis of eight cohort studies including 8669 participants. Alzheimers Dement 7: 412–417. pmid:21784352
  6. 6. Ho RC, Cheung MW, Fu E, Win HH, Zaw MH, et al. (2011) Is high homocysteine level a risk factor for cognitive decline in elderly? A systematic review, meta-analysis, and meta-regression. Am J Geriatr Psychiatry 19: 607–617. pmid:21705865
  7. 7. Obeid R, Herrmann W (2006) Mechanisms of homocysteine neurotoxicity in neurodegenerative diseases with special reference to dementia. FEBS Lett 580: 2994–3005. pmid:16697371
  8. 8. Gisslen M, Hagberg L, Brew BJ, Cinque P, Price RW, et al. (2007) Elevated cerebrospinal fluid neurofilament light protein concentrations predict the development of AIDS dementia complex. J Infect Dis 195: 1774–1778. pmid:17492593
  9. 9. CDC (1992) 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep 41: 1–19.
  10. 10. Hagberg L, Cinque P, Gisslen M, Brew BJ, Spudich S, et al. (2010) Cerebrospinal fluid neopterin: an informative biomarker of central nervous system immune activation in HIV-1 infection. AIDS Res Ther 7: 15. pmid:20525234
  11. 11. Magera MJ, Lacey JM, Casetta B, Rinaldo P (1999) Method for the determination of total homocysteine in plasma and urine by stable isotope dilution and electrospray tandem mass spectrometry. Clin Chem 45: 1517–1522. pmid:10471655
  12. 12. Bongiovanni M, Casana M, Pisacreta M, Tordato F, Cicconi P, et al. (2007) Predictive factors of hyperhomocysteinemia in HIV-positive patients. J Acquir Immune Defic Syndr 44: 117–119. pmid:17195739
  13. 13. Look MP, Riezler R, Berthold HK, Stabler SP, Schliefer K, et al. (2001) Decrease of elevated N,N-dimethylglycine and N-methylglycine in human immunodeficiency virus infection during short-term highly active antiretroviral therapy. Metabolism 50: 1275–1281. pmid:11699044
  14. 14. Smulders YM, de Man AM, Stehouwer CD, Slaats EH (1998) Trimethoprim and fasting plasma homocysteine. Lancet 352: 1827–1828.
  15. 15. Smulders YM, Spoelstra-de Man AM, Slaats EH, Weigel HM, Stehouwer CD, et al. (2001) Trimethoprim-sulphamethoxazole as primary Pneumocystis carinii prophylaxis does not increase serum homocysteine levels in HIV-positive subjects. Eur J Intern Med 12: 363–365. pmid:11395300
  16. 16. Schroecksnadel K, Frick B, Winkler C, Leblhuber F, Wirleitner B, et al. (2003) Hyperhomocysteinemia and immune activation. Clin Chem Lab Med 41: 1438–1443. pmid:14656023
  17. 17. Smith I, Howells DW, Kendall B, Levinsky R, Hyland K (1987) Folate deficiency and demyelination in AIDS. Lancet 2: 215.
  18. 18. Tan SV, Guiloff RJ (1998) Hypothesis on the pathogenesis of vacuolar myelopathy, dementia, and peripheral neuropathy in AIDS. J Neurol Neurosurg Psychiatry 65: 23–28. pmid:9667556
  19. 19. Fuchs D, Jaeger M, Widner B, Wirleitner B, Artner-Dworzak E, et al. (2001) Is hyperhomocysteinemia due to the oxidative depletion of folate rather than to insufficient dietary intake? Clin Chem Lab Med 39: 691–694. pmid:11592434
  20. 20. Frick B, Gruber B, Schroecksnadel K, Leblhuber F, Fuchs D (2006) Homocysteine but not neopterin declines in demented patients on B vitamins. J Neural Transm 113: 1815–1819. pmid:16988797
  21. 21. Savage DG, Lindenbaum J, Stabler SP, Allen RH (1994) Sensitivity of serum methylmalonic acid and total homocysteine determinations for diagnosing cobalamin and folate deficiencies. Am J Med 96: 239–246. pmid:8154512
  22. 22. Remacha AF, Cadafalch J, Sarda P, Barcelo M, Fuster M (2003) Vitamin B-12 metabolism in HIV-infected patients in the age of highly active antiretroviral therapy: role of homocysteine in assessing vitamin B-12 status. Am J Clin Nutr 77: 420–424. pmid:12540403
  23. 23. Beach RS, Morgan R, Wilkie F, Mantero-Atienza E, Blaney N, et al. (1992) Plasma vitamin B12 level as a potential cofactor in studies of human immunodeficiency virus type 1-related cognitive changes. Arch Neurol 49: 501–506. pmid:1580812
  24. 24. Kieburtz KD, Giang DW, Schiffer RB, Vakil N (1991) Abnormal vitamin B12 metabolism in human immunodeficiency virus infection. Association with neurological dysfunction. Arch Neurol 48: 312–314. pmid:1848071
  25. 25. Robertson KR, Stern RA, Hall CD, Perkins DO, Wilkins JW, et al. (1993) Vitamin B12 deficiency and nervous system disease in HIV infection. Arch Neurol 50: 807–811. pmid:8352665