Changes in PINCH levels in the CSF of HIV+ individualscorrelate with hpTau and CD4 count

Radhika Adiga & Ahmet Y. Ozdemir &

Alexandra Carides & Melissa Wasilewski & William Yen &

Pallavi Chitturi & Ronald Ellis & Dianne Langford

Received: 21 January 2014 /Revised: 31 March 2014 /Accepted: 4 April 2014# Journal of NeuroVirology, Inc. 2014

Abstract Several studies report associations between the par-ticularly interesting new cysteine histidine-rich (PINCH) pro-tein and HIV-associated CNS disease. PINCH is detected inthe CSF of HIV patients, and changes in levels during diseasemay be indicative of changes in disease status over time.PINCH binds hyperphosphorylated Tau (hpTau) in the brainand CSF, but little is known about the relevance of theseinteractions to HIV CNS disease. In this study, PINCH andhpTau levels were assessed in three separate CSF samplescollected longitudinally from 20 HIV+ participants before andafter initiating antiretroviral therapy or before and after achange in the treatment regimen. The intervals were approx-imately 1 (T2) and 3–7 (T3) months from the initial visit(baseline, T1). Correlational analyses were conducted forCSF levels of PINCH and hpTau and other variables includingblood CD4 T-cell count, plasma and CSF viral burden, CSFneopterin, white blood cell (WBC) count, and antiretroviralCNS penetration effectiveness (CPE). Values for PINCH andhpTauwere determined for each patient by calculating the foldchanges between the second (T2) and third measurements(T3) from the baseline measurement (T1). Statistical analysesshowed that the fold changes in CSF PINCH protein from T1to T2 were significantly higher in participants with CD4

counts >200 cells/mm3 at T2 compared to those with CD4counts <200 cells/mm3 at T2. This trend persisted irrespectiveof plasma or CSF viral burden or antiretroviral therapy CPEscores. The fold changes in PINCH levels between T1 and T2,and T1 and T3 were highly correlated to the fold changes inhpTau at T2/T1 and T3/T1 (correlation coefficient=0.69,p<0.001; correlation coefficient=0.83, p<0.0001, respective-ly). In conclusion, in these HIV participants, changes in CSFlevels of PINCH appear to correlate with changes in bloodCD4 count and with changes in CSF hpTau levels, but notwith plasma or CSF viral burden, neopterin, WBC, or antire-troviral regimen CPE.

Keywords CD4 count . hpTau . PINCH . HIV . CNS .

CSF

Introduction

Particularly interesting new cysteine histidine-rich (PINCH)protein is an adaptor protein involved in cytoskeletal organi-zation, cell attachment, and survival (Wu 1999; Rearden1994; Tu et al. 2001), and is suggested to play an importantrole in neurodegenerative diseases including human immuno-deficiency virus encephalitis (HIVE) and Alzheimer’s disease(AD) (Rearden et al. 2008; Jatiani et al. 2010; Ozdemir et al.2013). PINCH1 protein consists of five LIM domains and hasno known catalytic activity (Wu 1999; Rearden 1994; Wu2004). The expression of PINCH1 is essential during devel-opment for cell proliferation and migration, to maintain neu-ronal polarity and synaptodendritic connections, and knock-out is embryonically lethal (Guo et al. 2007; Li et al. 2005).While PINCH is expressed at high levels during developmentand in disease, in healthy patients, PINCH is nearly undetect-able. In the brain, PINCH is upregulated in dystrophic neuronsand is also present in brain parenchyma with no apparent

R. Adiga :A. Y. Ozdemir :M. Wasilewski :W. Yen :D. Langford (*)School of Medicine, Department of Neuroscience, TempleUniversity, 3500 N. Broad Street, MERB 750, Philadelphia,PA 19140, USAe-mail: tdl@temple.edu

A. Carides : P. ChitturiCenter for Statistical Analysis, Department of Statistics, Fox Schoolof Business, Temple University, Philadelphia, PA, USA

R. EllisDepartment of Neurosciences, University of California San DiegoSchool of Medicine, La Jolla, CA, USA

J. Neurovirol.DOI 10.1007/s13365-014-0252-8

association with cellular components (Rearden et al. 2008;Jatiani et al. 2010).

Hyperphosphorylation of Tau results in Tau’s dissociationfrom microtubules and mislocalization to the neuronal somaand dendrites. Accumulation of hyperphosphorylated Tau(hpTau) is a common pathological feature of AD and isreported in HIV, as well (Ozdemir et al. 2013; Anthony et al.2006; Patrick et al. 2011). In this context, our recent data showthat PINCH binds to hpTau in HIV and AD patients’ brainsand loses solubility along with hpTau (Ozdemir et al. 2013).Moreover, we have shown that the HIV protein Tat and thechemokine TNF-α induce both PINCH expression and hpTauin human primary neurons in vitro (Jatiani et al. 2010). How-ever, the significance of the presence of PINCH in the CSF ofHIV patients is currently unclear.

The relevance of plasma and CSF viral loads and CD4count in HIV-associated neuropathology, prognosis, and re-sponse to treatment has been extensively studied (Fox 2013;Marcotte et al. 2013; Haughey et al. 2013; Megra et al. 2013).Numerous studies have addressed potential HIV-associatedCSF biomarkers that may be used alone or in combinationto either predict CNS disease severity or to predict the pro-gression of HIV CNS disease (Brew et al. 2005; Anderssonet al. 1999; Ellis et al. 1998; Green et al. 2000; Fields et al.2013). Similarly, many studies have assessed changes inhpTau levels in CSF as a biomarker in AD, and more recentlyin HIV (Ciborowski 2009). Although results from some re-ports are conflicting regarding CSF levels of Tau protein inHIV, most agree that elevations in CSF inflammatory factorspersist throughout disease. For example, a recent review fromPrice et al. describes the value of CSF assessment to improveour understanding of the evolution of HIV neuropathology asit relates to changes in immune system function (Price et al.2013). Additionally, in AD, changes in neurodegenerativemarkers are reported to precede structural changes in the brainthat are detectable by MRI or clinically as alterations incognition (Lazarczyk et al. 2012). Therefore, reliable neuro-degenerative and neuroinflammatory biomarkers may be use-ful in predicting changes in disease progression in HIV-associated neurocognitive disorders (HAND), AD, and otherneurodegenerative diseases and may influence treatment plansfor managing patients.

Increased life expectancies of HIV patients have led to atremendous growth in the population of older individualsliving with HIV. While combination antiretroviral therapy(CART) is one of the major factors contributing to this phe-nomenon, both HIV infection and CART accelerate the agingprocess (Nath 2012; Holt et al. 2012; Bhatia et al. 2012).Antiretroviral therapy causes a substantial decrease in immuneactivation and an increase in CD4 count, but recovery in thesepatients is still characteristic of immune senescence and isdependent in part on both CD4 nadir and age at CARTinitiation (Deeks 2011; Kitchen et al. 2011).

Several studies report the early development of an “elderlyimmune system profile” in HIV patients resulting in a deplet-ed naïve T cell population, lymphoid tissue fibrosis, andinverted CD4 to CD8 ratio (for review, see (Jefferys 2013)).Also, shortened telomeres in Tcells are reported in both HIV+patients and in aging HIV-negative adults (Leeansyah et al.2013; Tassiopoulos et al. 2012; Pathai et al. 2014). Further-more, MRI points to immune suppression as a key player inbrain volume reduction in clinically stable HIV patients onCART, correlating with lower CD4 nadir (Hua et al. 2013).These studies suggest the significance of the immune systemin accelerating aging in HIV patients.

Despite CART adherence, some patients do not have afavorable response as evidenced by persistent virus in plasmaand CSF and low blood CD4 counts. In this context, immu-nosuppressed HIV patients including immunological non-responders have increased morbidity and mortality comparedto HIV patients with more favorable immune responses (Hunt2012); whereas, HIV patients with substantial immune recov-ery and optimal immune system functioning have life expec-tancies similar to those of HIV-negative individuals (Hunt2012). Hence, longitudinal studies and clinical trials exploringbiomarkers to examine the relationship between immune sys-tem and HIV are essential to effectively evaluate age-relateddisease CNS progression and treatment efficacy in HIV pa-tients. Earlier studies report that although PINCH CSF levelsare significantly elevated in all HIV patients, CSF levels ofPINCH are significantly higher in HIV patients without en-cephalitis compared to those with HIVE (Rearden et al. 2008),suggesting that changes in PINCH CSF levels may relate inpart to inflammatory status of the brain.

In this study, we assessed changes in PINCH levels in acohort of HIV study participants from whom at least threeCSF and blood samples were collected over 3–7 months. Ourfindings show that as CD4 count increases, PINCH levels inthe CSF increase, as determined by increased levels from onetime point to the next. Likewise, changes in levels of CSFPINCH correlate with changes in CSF hpTau levels, but donot correlate with plasma or CSF viral burden. Increasedunderstanding of the pathophysiology of HIV infection ofthe CNS in the presence of CART, changes in viral load, andimmune system status may improve clinical management andtreatment strategies for HIV patients with neurologicaldisorders.

Methods

Study design and participants

Temple University and UCSDHuman Subjects IRB approvedthe study, and a written informed consent was obtained fromall study participants. Longitudinal CSF and blood samples

J. Neurovirol.

were collected from 20 ambulatory HIV+ study volunteersbetween 1997 and 2004 enrolled at the HIV NeurobehavioralResearch Center, University of California San Diego, SanDiego, CA. Clinical and laboratory data from CSF and bloodsamples for study participants are presented in Table 1. Meanage of participants was 39 years (range, 27–54 years) andincluded 12 Caucasians, 5 African-Americans, 1 Hispanic,and 1 participant of undetermined background. All studyparticipants received nucleoside or nucleotide reverse tran-scriptase inhibitors (NRTI), 12 participants received non-nucleoside RTI, and 11 participants received proteaseinhibitors.

For each participant, at least three samples were collectedwith the first draw representing baseline, T1. The second draw(T2) ranges from 27–50 days from T1 with a median of 33days, and 29 days representing the first and 36 daysrepresenting the third quartile. Likewise, the third draw (T3)ranges from 60–207 days from T1 with a median of 96 days,and 85 days representing the first and 123 days representingthe third quartile (Table 1). Baseline (T1) CSF and blooddraws corresponded to either initiation of CART or a changein treatment regimen due to failure as indicated by one or moreof the following parameters: increased plasma and/or CSFviral burden, plasma or CSF viral burden going from unde-tectable to detectable levels, or a significant decrease in CD4count or a failure of CD4 count to improve over time. Plasmaand CSF HIV viral load, blood CD4 count, and CSF whiteblood cell (WBC) count were obtained for most of the partic-ipants at each visit. Classification of participants as virologicresponders (n=17) or non-responders (n=3) was based on one

log reduction of the plasma viral load from baseline (T1) atsubsequent collections (T2 or T3).

No history of CNS opportunistic infections including cy-tomegalovirus, toxoplasmosis, or JC virus with progressivemultifocal encephalopathywas reported for study participants.Self-reported current and long-term substance abuse and de-pendence history for alcohol, cocaine, and cannabis wereobtained at the time of each visit for nine participants.

CSF and blood measurements

Plasma and CSF HIV RNA levels (log10 copies/ml) werequantified using the Amplicor HIV-1 Monitor test with Ultra-sensitive Specimen Preparation (Roche Molecular Systems,Pleasanton, CA). The limit of detection was 50 copies/ml forCSF and 200 copies/ml for plasma, and suppression of HIVRNA was defined as a decline in levels from detectable toundetectable (Marra et al. 2009).

CPE scores for antiretroviral medications

CNS penetration effectiveness (CPE) ranks were assigned toeach component of CART as previously described (Letendreet al. 2010). The CPE rank was derived by dividing the sum ofranks assigned for each antiretroviral drug included in thepatient regimen by the number of medications. Scores rangedfrom 1 to 4 with a higher score representing more effectivepenetration.

Sample processing and Western blot analysis

CSF samples were stored at −80 °C, thawed on ice, andcentrifuged briefly at top speed at 4 °C to remove cell debris.An equal volume (4 μl) of each CSF sample was loaded perwell onto 4–12% Bis-Tris pre-cast gels (Invitrogen, Carlsbad,CA, USA). Samples were separated by electrophoresis andtransferred onto nitrocellulose membranes (Bio-Rad). Mem-branes were blocked in 5 % non-fat milk or 5 % BSA in Tris-buffered saline with 0.1 % Tween 20 (1X TBST) for 30 minbefore overnight incubation at 4 °C with primary antibodies.Primary antibodies included PINCH1 (1:1,000, BD, Rock-ville, MD, USA), anti-Tau antibodies, S262 (1:1000, Abcam),and S396 (1:1,000, Abcam). Membranes were washed threetimes in 1X TBST and incubated in secondary anti-mouseIgG2a or anti-rabbit HRP (1:10,000, Thermo Scientific, Lafa-yette, CO) antibody for 1 h at room temperature and visualizedwith enhanced chemiluminescence (ECL) prime (AmershamPharmacia Biotech, Piscataway, NJ, USA). Densitometry wasperformed using 3D ImageJ software (Rasband WS: ImageJ.In., 1.37 edn. Bethesda: NIH 1997). To assess the potentialcorrelation of PINCH with hpTau, CD4 count, CSF andplasma HIV viral burden, and CSF WBC count at each timepoint, fold change of PINCH was calculated as the ratio of the

Table 1 Summary statistics representing the median and interquartilerange for plasma HIV RNA load, CSF HIV RNA load, CSF white bloodcell count and blood CD4 count

Sample collection Median (interquartile range)

Plasma HIV RNA (log10RNA copies/ml)

First time point (baseline, T1) 4.92 (4.71–5.54)

Second time point (T2) 2.99 (2.54–3.2)

Third time point (T3) 2.30 (2.30–2.53)

CSF HIV RNA (log10RNA copies/ml)

First time point (baseline, T1) 3.54 (2.82–3.86)

Second time point (T2) 1.95 (1.73–2.26)

Third time point (T3) 1.69 (1.69–1.69)

CSF white blood cell count (cells/mm3)

First time point (baseline, T1) 3 (1–12)

Second time point (T2) 3 (2–4)

Third time point (T3) 2 (1–3)

CD4 count (cells/mm3)

First time point (baseline, T1) 139 (67–237)

Second time point (T2) 285 (183–375)

Third time point (T3) 328 (233–362)

J. Neurovirol.

densitometry value for the second time point (T2) and the firsttime point (baseline, T1) represented as T2/T1. Similarly, T3/T1 represents the fold change calculated as the ratio of thedensitometry value for the third (T3) time point and the first(T1) time point.

CSF Neopterin ELISA

CSF neopterin concentrations were analyzed according tomanufacturer’s instructions using a commercially availablesensitive enzyme-linked immunosorbent assay (NeopterinELISA, IBL international GMBH, Hamburg, Germany). Cu-bic spline interpolation with a not-a-knot condition was usedto calculate neopterin concentrations from optical densityvalues of duplicate samples (http://www.akiti.ca/CubicSpline.html) (Hautala et al. 2013).

Statistical analysis

Median values with interquartile range for plasma HIV RNAlog, CSF HIV RNA log, CSFWBC count, and CD4 count forthe three time points were computed to describe the distribu-tion of the data. Kruskal-Wallis one-way analysis of variance(ANOVA) test was conducted to assess the difference in meanvalues for PINCH across the time points and to assess thedifference in mean values for PINCH between responders andnon-responders. Spearman correlation coefficient was com-puted for PINCH and hpTau (S262 and S396) in study partic-ipants. Spearman correlation coefficients were also calculatedto assess correlation between fold change of PINCH and HIVRNA values in CSF and plasma, as well as between PINCHand CSF WBC count. Similarly, Spearman correlation coeffi-cients were calculated to assess correlation between CD4count and fold change of PINCH, and between CD4 countand S262, and S396. The α value for all the statistical testswas retained at 0.05. Statistical analyses were conducted usingPrism 6 (GraphPad Software, Inc.).

Results

Changes in PINCH levels in HIV participants’ CSF over time

To assess CSF levels of PINCH after the initiation of CART orafter a change in the CART regimen, samples from three differ-ent time points per patient were assessed by Western analysis.Distinct PINCH bands were detected at approximately 55 and71 kDa (Fig. 1a), as previously reported (Rearden et al. 2008;Ozdemir et al. 2013). A third PINCH immunoreactive band wasdetected at approximately 84 kDa in some cases, reminiscent ofthe approximate size detected by immunoprecipitation ofPINCH and hpTau (Ozdemir et al. 2013). Detection of multiplePINCH immunoreactive bands has been previously reported,

but currently, the significance of the different bands is unknown.To determine if PINCH levels within a given participant differedover time, we compared levels of PINCH from the three differenttime points for each participant as shown by a representative blotfor one virologic non-responder and one responder patient(Fig. 1a). When we compared the changes at T2 from baselinefor all participants, in 10 of the 17 virologic responder participantsand 1 non-responder participant, PINCH levels increased abovebaseline or remained unchanged at T2 (Fig. 1b, blue and blackcircles, respectively). Changes were statistically non-significant(p=0.2196). In the remaining seven responder participants (yellowcircles) and two non-responder participants (pink circles), PINCHlevels were below baseline at T2 (Fig. 1b). This general trendwasalso observed for T3/T1 but changes were more variable.

Changes in hpTau levels in HIV patients’ CSF followeda similar trend as PINCH levels

Changes in hp Tau levels in the CSF of HIV+ participantswere assessed by Western analyses for each of the three

Fig. 1 PINCH levels in the CSF of HIV participants. a RepresentativeWestern blot of PINCH in CSF from two different HIV+ participants, onenon-responder (NR) and one responder (R). Western blot of CSF collectedat three different time points and reacted with anti-PINCH antibodydetects bands at molecular weights of approximately 55, 71, or 84 kDa.T1 (baseline), T2, and T3 correspond to the first, second, and third timepoints of sample collection, respectively. b Kruskal-Wallis one-wayANOVA of PINCH comparing fold changes to baseline values in eachpatient yielded statistically non-significant results. T2/T1 and T3/T1 rep-resent fold change from the baseline value (T1). Blue circles representresponder HIV+ participants with increased fold change of PINCH at T2or T3 compared to T1. Yellow circles represent responder HIV partici-pants with decreased fold change of PINCH at T2 compared to T1. Theblack circle represents a non-responder HIV+ participant with an in-creased fold change of PINCH at T2 or T3 compared to T1. Pink circlesrepresent non-responder HIV+ participants with decreased fold change ofPINCH at T2 or T3 compared to T1

J. Neurovirol.

collection time points. A representativeWestern blot from onevirologic non-responder (NR) and one responder participant(R) shows the typical banding pattern for these participants(Fig. 2a). Changes in hpTau levels in approximately 50 % ofresponders (n=9, blue squares) and one non-responder (blacksquare) increased or remained unchanged at the second timepoint (T2) (Fig. 2b). Likewise, hpTau levels decreased in eightvirologic responders (yellow squares) and two non-responders(pink squares) (Fig. 2b). Fold changes in hpTau at the thirdtime point (T3) followed a similar trend with eight responders’levels (blue squares) and one non-responder’s (black square)level at T2, increasing or remaining unchanged (Fig. 2b).Changes were non-significant (p= 0.4467).

PINCH levels correlate with hpTau levels in CSFof HIV + patients

To assess if changes in levels of PINCH and hpTau correlatewith one another, fold changes in PINCH and hpTau weredetermined by Western analyses by calculating the ratio of thedensitometry value for the second time point (T2) to baseline

levels and the third time point (T3) to baseline. Spearmancorrelation coefficient analyses indicated that the fold changein PINCH levels between the first and second time points washighly correlated to the hpTau variable (correlation coeffi-cient=0.70, p=0.001) (Fig. 3a). Fold changes in PINCH levelsbetween the first and third time points correlated with hpTau,as well (correlation coefficient=0.83, p<0.0001) (Fig. 3b).There was no correlation between the second and third PINCHratios with hpTau (data not shown). Although the trendpersisted, analyses for the correlation between PINCH andhpTau at S396 showed no statistical significance (data notshown). These findings suggest a relationship between the foldchanges in PINCH and hpTau in HIV participants on CART.

Changes in PINCH CSF levels correlate with CD4 countin HIV patients

To assess if changes in CSF PINCH levels correlate with otherclinical parameters for this group of study participants,PINCH fold changes were compared to CD4 count, plasma

Fig. 2 hpTau levels in CSF of HIV+ participants. a RepresentativeWestern blot of hpTau (S262) in CSF from two different HIV+ partici-pants, one non-responder (NR) and one responder (R). Western blot ofCSF collected at three different time points and reacted with anti-S262hpTau antibody detects bands at an approximate molecular weight of55 kDa. T1 (baseline), T2, and T3 correspond to the first, second, andthird time points of sample collection. b Kruskal-Wallis one-wayANOVA of S262 comparing fold changes to baseline values in eachpatient yielded statistically non-significant results. T2/T1 and T3/T1 rep-resent fold change from the baseline value (T1). Blue squares representresponder HIV+ participants with increased fold change of hpTau at T2 orT3 compared to T1. Yellow squares represent responder HIV+ partici-pants with decreased fold change of hpTau at T2 or T3 compared to T1.The black squares represent non-responder HIV+ participants with anincreased fold change of hpTau at T2 or T3 compared to T1. Pink squaresrepresent non-responder HIV+ participants with decreased fold change ofhpTau at T2 or T3 compared to T1

Fig. 3 Changes in PINCH levels in CSF correlate with changes in hpTaulevels. a Fold change in PINCH (y-axis) compared to hpTau (x-axis) frombaseline (T1) to T2 or b from baseline (T1) to T3. Blue circles representresponder HIV+ participants with increased fold change of PINCH at T2or T3 compared to T1. Yellow circles represent responder HIV+ partici-pants with decreased fold change of PINCH at T2 or T3 compared to T1.The black circle represents a non-responder HIV+ participant with anincreased fold change of PINCH at T2 or T3 compared to T1. Pinkcircles represent non-responder HIV+ participants with decreased foldchange of PINCH at T2 or T3 compared to T1. Spearman correlationcoefficients for PINCH ratios and hpTau (S262) at a T2/T1, p=0.001and b T3/T1, p<0.0001

J. Neurovirol.

and CSF viral RNA, CSF WBC counts, and neopterin levels.Spearman correlation coefficients showed no correlation be-tween PINCH and HIV RNA, between PINCH and WBC, orbetween PINCH and CPE scores (not shown). As previouslyreported by Price et al., analyses from a subset of theseparticipants confirmed that CSF neopterin levels and CSFviral load correlated. However, neopterin levels did not corre-late with CSF PINCH (data not shown).

On the other hand, Spearman correlation coefficients forCD4 count and PINCH indicated significant correlation (cor-relation coefficient=0.72, p=0.0067) (Fig. 4a, circles), andCD4 count and hpTau at S262 (squares) yielding a correlationcoefficient=0.59 and p=0.0381 (Fig. 4a). As shown in

Fig. 4a, PINCH levels increased or remained unchanged inall participants with CD4 counts over 200 cells/mm3, whereasall participants whose CD4 count dropped below 200 cells/mm3 had less PINCH in the CSF (yellow and pink circles).With regard to hpTau levels, this trend was observed for all butone participant (Fig. 4a, yellow square).

Taken together, these results show that changes in PINCHCSF levels correlate with changes in CD4 count (Fig. 4b).Moreover, our data show that increased CSF levels of PINCHcorrespond with a more favorable immune response as de-fined by improved CD4 count. Previous data reported thatincreased PINCH is observed in the brains and CSF of HIVpatients, but it was also shown that patients with HIVE had

Fig. 4 Changes in PINCH in CSF correlate with CSF levels of hpTauand blood CD4 count. a T1 and T2 correspond to the first and second timepoints of sample collection. T2/T1 represents fold change from thebaseline value (T1). Blue represents responder HIV+ participants withincreased fold change of PINCH (circles) or hpTau (squares) at T2compared to T1. Yellow represents responder HIV+ participants withdecreased fold change of PINCH (circles) or hpTau (squares) at T2compared to T1. Pink represents non-responder HIV+ participants withdecreased fold change of PINCH (circles) or hpTau (squares) at T2compared to T1. Black represents a non-responder HIV+ participant with

an increased fold change of PINCH (circle) and hpTau (square) at T2compared to T1. Circles represent PINCH, and squares represent hpTau(S262). Spearman correlation coefficient fold change of PINCH at T2compared to T1 and CD4 count at T2 (p=0.0067) and fold change ofhpTau (S262) at T2 compared to T1 and CD4 count at T2 (p=0.0381). bVenn diagram depicting the correlations between fold change of PINCHat T2 compared to T1 and CD4 count at T2 in responders and non-responders. The blue circle represents responders, the pink circle repre-sents non-responders, and the intersecting yellow region represents re-sponders with decreased PINCH ratio and CD4 count <200 cells/mm3

J. Neurovirol.

more PINCH in their brains compared to CSF (Rearden et al.2008). Likewise, HIV patients with no CNS alterations hadless PINCH in the brain and more in the CSF. These findingsare supported by our data showing that as CD4+ count in-creases, PINCH levels in the CSF increase as well.

Discussion

In the current study, we addressed the potential significance ofPINCH in the CSF of HIV+ individuals. Clinical and labora-tory information including age, blood CD4 count, CSF andplasma viral load, CSF WBC count, and CPE scores ofantiretroviral therapy were compared with fold changes inCSF PINCH and hpTau levels. Our results suggest that chang-es in CSF levels of PINCH correlate with CSF hpTau levels inthis cohort of HIV participants. Previous studies reported thatPINCH and hpTau bind one another both in vitro, in vivo, andin the brains and CSF of HIV patients (Ozdemir et al. 2013).Our earlier studies also reported higher levels of PINCH in thebrain and CSF in both HIV with no CNS alterations andHIVE, compared to HIV-negative adults. However, PINCHlevels in CSF in HIV patients with no CNS alterations weresubstantially greater than that in HIVE. However, studies withmatched brain tissue and CSF samples fromHIV+ individualsare required to validate and understand the significance ofthese observations. Similarly, alterations in the solubility ofPINCHwere also associated with solubility changes in hpTau.These studies support previous findings and suggest a corre-lation between PINCH and hpTau levels in HIV-associatedCNS disease. Although the mechanisms underlying this asso-ciation are unknown, there are several points of intersectionbetween PINCH and Tau pathways that may contribute totheir interaction.

One possible contributing mechanism is through PINCH’sbinding partners that are involved in hyperphosphorylation ofTau and rely on interacting with PINCH to maintain theircatalytic activity. For example, the most well-characterizedbinding partner of PINCH is integrin-linked kinase (ILK).ILK is an ankyrin repeat containing serine-threonine kinasethat phosphorylates multiple kinases such as GSK3β andAKT involved in aberrant hyperphosphorylation of Tau(Legate et al. 2006). In the absence of PINCH, ILK activityis diminished; thus, alterations in levels of PINCH may con-tribute to changes in formation of hpTau. Likewise, PINCHbinds to and inactivates protein phosphatase 1α (PP1α) that isinvolved in the dephosphorylation of Tau residues essentialfor normal Tau-microtubule interactions (Eke et al. 2010). Theheat shock response represents a third potential pathway thatmay contribute to PINCH/hpTau interactions, as we haveshown that PINCH binds to both hpTau and to the E3 ubiq-uitin ligase (CHIP) (Ozdemir et al. 2013). In this context,knockdown of PINCH in neurons where hpTau was induced

resulted in less hpTau detection, which could be a result of lessformation of hpTau, more efficient clearance, or both. Thepotential interaction(s) among these pathways and PINCHand Tau levels in the CSF are currently unknown; however,changes in binding dynamics and release from the brain intothe CSF is an area warranting more investigation.

A second significant and unexpected finding was that thefold changes in PINCH levels over baseline correlated withblood CD4+ count. Interestingly, all participants with a CD4+count above 200 cells/mm3 at T2 showed increased or nochange in PINCH CSF levels and participants whose CD4count decreased below 200 had less PINCH in the CSF. BloodCD4+ count is used as one of the clinical measures of HIVinfection, in opportunistic infections in HIV, and as a predictorof response to CART. A CD4 count less than 200 cells/mm3 isone of the diagnostic markers for acquired immunodeficiencysyndrome. The results from our current study might representan important correlation between PINCH CSF levels andsystemic immune response in HIV. The connection betweenimproved immune response and greater levels of PINCH inthe CSF is unclear. But as pointed out in earlier studies, HIVpatients with less severe CNS alterations may in fact clearboth PINCH and hpTau from the brain more efficiently,resulting in the detection of more PINCH/hpTau in the CSFrelative to patients with increased hpTau accumulation in thebrain and, by virtue of association, more PINCH in the brainas well.

About 20 % of HIV patients are immunological non-responders with low blood levels of immune cells, increasedimmune activation, immunosenescence, and apoptosis even inthe presence of effective antiretroviral therapy and significantviral suppression. Various causes for poor immunologicalresponse include older age, ineffective antiretroviral therapy,and low CD4 nadir with a CD4 count <200 cells/mm3. Interest-ingly, the two virologic responders in our study with a CD4count <200 cells/mm3 and a decrease in PINCH ratios wereover 45 years of age; whereas, others were younger than40 years of age, suggesting that age may be a factor in CD4count and PINCH levels.

There were several limitations in this study. Although thisstudy was longitudinal in some aspects, the collection datesamong participants for each time point were not conducted atthe exact same intervals. Rather, collection points for T2 andT3 were 33 (29–36) and 96 (85–123) days, respectively,representing the median (interquartile range with the firstand third quartile). This rather wide range may account forthe variation observed at T3 in PINCH and hpTau (Figs. 1band 2b). Of note, changes in PINCH levels at 1–1.5 monthsfrom baseline showed stronger correlations with CD4. Anoth-er limitation in the study is that we were unable to considerCD4 nadir in our correlation analyses since this data was notavailable for all patients. Because CD4 nadir is a better clinicaland diagnostic measure of immune system response to HIV

J. Neurovirol.

infection and treatment, future studies will consider the CD4nadir to increase our understanding of the relationship be-tween CSF PINCH and immune system. Thus, future studiesto understand the potential significance of PINCH in brainsand CSF of HIV patients should be expanded to include manymore patients and consider age, immune response, and re-sponse to treatment.

In this context, in our current study, we observed a trend ofdecreased fold change for PINCH in older HIV participants inthe responder treatment group with CD4 count below200 cells/mm3, suggesting PINCH protein’s association withaging in defining disease pathology. Thus, future studiesexploring the effects of aging in HIV patients and levels ofPINCH in the brain and CSF in response to CARTare needed.Furthermore, CD4+/CD8+ T-cell ratios in CSF and bloodshould also be analyzed for assessing the response to treat-ment, effect of immunosenescence in HIV, and disease pa-thology in the context of PINCH expression levels. Under-standing the role of PINCH in CNS in HIV infection will helpbetter define the function of PINCH in the pathophysiology ofaberrant Tau formation in HIV, in immune response to infec-tion and therapy, and the significance of its detection in CSFand brain in HIV, and perhaps AD, and other neurodegenerativediseases. In conclusion, PINCH CSF levels appear to correlatewith immune responses in HIV participants, suggesting anotherrole for PINCH in CNS disease pathology including in HIVandother neurodegenerative diseases.

Acknowledgments This work was supported byNIMH085602 to TDLand NIMH R0158076 to RE.

Conflict of interest The authors Radhika Adiga, Ahmet Y. Ozdemir,Alexandra Carides, Melissa Wasilewski, William Yen, Pallavi Chitturi,Ronald Ellis, and Dianne Langford declare that they have no conflicts ofinterest.

References

Andersson L, Blennow K, Fuchs D, Svennerholm B, Gisslen M (1999)Increased cerebrospinal fluid protein tau concentration in neuro-AIDS. J Neurol Sci 171(2):92–96

Anthony IC, Ramage SN, Carnie FW, Simmonds P, Bell JE (2006)Accelerated Tau deposition in the brains of individuals infected withhuman immunodeficiency virus-1 before and after the advent ofhighly active anti-retroviral therapy. Acta Neuropathol 111(6):529–538

Bhatia R, Ryscavage P, Taiwo B (2012) Accelerated aging and humanimmunodeficiency virus infection: emerging challenges of growingolder in the era of successful antiretroviral therapy. J Neurovirol18(4):247–255

Brew BJ, Pemberton L, Blennow K, Wallin A, Hagberg L (2005) CSFamyloid beta42 and tau levels correlate with AIDS dementia com-plex. Neurology 65(9):1490–1492

Ciborowski P (2009) Biomarkers of HIV-1-associated neurocognitivedisorders: challenges of proteomic approaches. Biomark Med 3(6):771–785

Deeks SG (2011) HIV infection, inflammation, immunosenescence, andaging. Annu Rev Med 62:141–155

Eke I, Koch U, Hehlgans S, Sandfort V, Stanchi F, Zips D, Baumann M,Shevchenko A, Pilarsky C, Haase M et al (2010) PINCH1 regulatesAkt1 activation and enhances radioresistance by inhibitingPP1alpha. J Clin Invest 120(7):2516–2527

Ellis RJ, Seubert P, Motter R, Galasko D, Deutsch R, Heaton RK, HeyesMP, McCutchan JA, Atkinson JH, Grant I (1998) Cerebrospinalfluid tau protein is not elevated in HIV-associated neurologic diseasein humans. HIV Neurobehavioral Research Center Group (HNRC).Neurosci Lett 254(1):1–4

Fields J, Dumaop W, Adame A, Ellis RJ, Letendre S, Grant I, Masliah E(2013) Alterations in the levels of vesicular trafficking proteinsinvolved in HIV replication in the brains and CSF of patients withHIV-associated neurocognitive disorders. J NeuroimmunePharmacol 8(5):1197–1209

Fox HS (2013) Biomarkers for NeuroAIDS: recent progress in the field. JNeuroimmune Pharmacol 8(5):1055–1058

Green AJ, Giovannoni G, Hall-Craggs MA, Thompson EJ, Miller RF(2000) Cerebrospinal fluid tau concentrations in HIV infected pa-tients with suspected neurological disease. Sex Transm Infect 76(6):443–446

Guo W, Jiang H, Gray V, Dedhar S, Rao Y (2007) Role of the integrin-linked kinase (ILK) in determining neuronal polarity. Dev Biol306(2):457–468

Haughey NJ, Zhu X, Bandaru VV (2013) A biological perspective ofCSF lipids as surrogate markers for cognitive status in HIV. JNeuroimmune Pharmacol 8(5):1136–1146

Hautala T, Partanen T, Sironen T, Rajaniemi S-M, Hautala N, Vainio O,Vapalahti O, Kauma H, Vaheri A (2013) Elevated cerebrospinalfluid neopterin concentration is associated with disease severity inacute puumala hantavirus infection. Clin Dev Immunol 1–4

Holt JL, Kraft-Terry SD, Chang L (2012) Neuroimaging studies of theaging HIV-1-infected brain. J Neurovirol 18(4):291–302

Hunt PW (2012) HIVand inflammation: mechanisms and consequences.Curr HIV/AIDS Reports 9(2):139–147

Jatiani A, Pannizzo P, Gualco E, Del-Valle L, Langford D (2010)Neuronal PINCH is regulated by TNF-alpha and is required forneurite extension. J Neuroimmune Pharmacol 6(3):330–340

Jefferys RaTH (2013) The immune system, HIV, & aging. In. Edited byBenzacar A. New York, New York: Treatment Action Group 1–25

Kitchen CM, Yeghiazarian L, Hoh R, McCune JM, Sinclair E, Martin JN,Deeks SG (2011) Immune activation, CD4+ T cell counts, andviremia exhibit oscillatory patterns over time in patients with highlyresistant HIV infection. PLoS One 6(6):e21190

Lazarczyk MJ, Hof PR, Constantin B, Giannakopoulos P (2012)Preclinical Alzheimer disease: identification of cases at risk amongcognitively intact olderindividuals. BMC Med 10:127–139

Leeansyah E, Cameron PU, Solomon A, Tennakoon S, Velayudham P,GouillouM, Spelman T, Hearps A, Fairley C, de Smit Vet al (2013)Inhibition of telomerase activity by human immunodeficiency virus(HIV) nucleos(t)ide reverse transcriptase inhibitors: a potential fac-tor contributing to HIV-associated accelerated aging. J Infect Dis207(7):1157–1165

Legate KR, Montanez E, Kudlacek O, Fassler R (2006) ILK, PINCH andparvin: the tIPP of integrin signalling. Nat Rev Mol Cell Biol 7(1):20–31

Letendre SL, Ellis RJ, Ances BM, McCutchan JA (2010) Neurologiccomplications of HIV disease and their treatment. Top HIV Med18(2):45–55

Li S, Bordoy R, Stanchi F, Moser M, Braun A, Kudlacek O, Wewer UM,Yurchenco PD, Fassler R (2005) PINCH1 regulates cell-matrix andcell-cell adhesions, cell polarity and cell survival during the peri-implantation stage. J Cell Sci 118(Pt 13):2913–2921

Marcotte TD, Deutsch R, Michael BD, Franklin D, Cookson DR, BhartiAR, Grant I, Letendre SL, Group C (2013) A concise panel of

J. Neurovirol.

biomarkers identifies neurocognitive functioning changes in hiv-infected individuals. J Neuroimmune Pharmacol 8(5):1123–1135

Marra CM, YZ, DBC, SL, SE, Henry K, RJE, BR, RWC, G et al (2009)Impact of combination antiretroviral therapy on cerebrospinal fluidHIV RNA and neurocognitive performance. AIDS (London,England) 23(11):1359–1366

Megra B, Eugenin E, Roberts T, Morgello S, Berman JW (2013) Proteaseresistant protein cellular isoform (PrP(c)) as a biomarker: clues intothe pathogenesis of HAND. J Neuroimmune Pharmacol 8(5):1159–1166

Nath A (2012) HIV and aging. J Neurovirol 18(4):245–246Ozdemir AY, Rom I, Kovalevich J, Yen W, Adiga R, Dave RS, Langford

D (2013) PINCH in the cellular stress response to tau-hyperphosphorylation. PloS one 8(3):e58232

Pathai S, Bajillan H, Landay LA, Kevin P (2014) High: Is HIVa model ofaccelerated or accentuated aging? J Gerontol A Biol Sci Med Sci68(10):doi:10.1093/gerona/glt1168

Patrick C, Crews L, Desplats P, Dumaop W, Rockenstein E, Achim CL,Everall IP, Masliah E (2011) Increased CDK5 expression in HIVencephalitis contributes to neurodegeneration via tau phosphoryla-tion and is reversed with Roscovitine. Am J Pathol 178(4):1646–1661

Price RW, Peterson J, Fuchs D, Angel TE, Zetterberg H, Hagberg L,Spudich S, Smith RD, Jacobs JM, Brown JN et al (2013) Approachto cerebrospinal fluid (CSF) biomarker discovery and evaluation inHIV infection. J Neuroimmune Pharmacol 8(5):1147–1158

Rasband WS: ImageJ. In., 1.37 edn. Bethesda: NIH; 1997

Rearden A (1994) A new LIM protein containing an autoepitope homol-ogous to “senescent cell antigen”. Biochem Biophys Res Commun201(3):1124–1131

Rearden A, Hurford RG, Luu N, Kieu E, Sandoval M, Perez-Liz G,Del Valle L, Powell H, Langford TD (2008) Novel expres-sion of PINCH in the CNS and its potential as a biomarkerfor HIV-associated neurodegeneration. J Neurosci Res 86(11):2535–2542

TassiopoulosK, LandayA, Collier AC, Connick E,Deeks SG,Hunt P, LewisDE, Wilson C, Bosch R (2012) CD28-negative CD4+ and CD8+ Tcells in antiretroviral therapy-naive HIV-infected adults enrolled in adultclinical trials group studies. J Infect Dis 205(11):1730–1738

Tu Y, Huang Y, Zhang Y, Hua Y, Wu C (2001) A new focal adhesionprotein that interacts with integrin-linked kinase and regulates celladhesion and spreading. J Cell Biol 153(3):585–598

Wu C (1999) Integrin-linked kinase and PINCH: partners in regulation ofcell-extracellular matrix interaction and signal transduction. J CellSci 112(Pt 24):4485–4489

Wu C (2004) The PINCH-ILK-parvin complexes: assembly, functionsand regulation. Biochim Biophys Acta 1692(2–3):55–62

Hua X, Boyle CP, Harezlak J, Tate DF, Yiannoutsos CT, Cohen R,Schifitto G, Gongvatana A, Zhong J, Zhug T, Taylor MJ, CampbelTB, Daar ES, Alger JR, Singer E, Buchthal S, Toga AW, Navia B,Thompson P (2013) the HIV Neuroimaging Consortium: Disruptedcerebral metabolite levels and lower nadir CD4+ counts are linked tobrain volume deficits in 210 HIV-infected patients on stable treat-ment. NeuroImage: Clinical 3:132–142

J. Neurovirol.